Revista Pădurilor 138(3) (2023) 001-044
Revista Pădurilor
disponibil online la: www.revistapadurilor.com
Revista tehnico-științifică editată de Societatea „Progresul Silvic"
COLEGIUL DE REDACȚIE
Redactor responsabil:
Prof. Dr. Ing- Stelian A. Borz
CUPRINS
Gabriel Osei Forkuo
O sinteză a metodelor tradiționale și moderne folosite în evaluarea
ergonomic-posturală............................................1
Membri:
Prof. Dr. Ing- loan V. Abrudan
Ing. Codruț Bîlea
Prof. Dr. Ing. Alexandru L. Curtu
Conf. Dr. Ing- Mihai Daia
Conf. Dr. Ing. Gabriel Duduman
Conf. Dr. Ing. Sergiu Horodnic
Elena Camelia Mușat
Domnul prcfesor dr. ing. Arcadie Ciubotarii împlinește 75 de
ani.........................................................35
Dan-Marian Gurean, Pantelimon Ularu, Victor Adrian Indreica
Prcf. dr. biol. Marius Danciu (1941-2023), un botanist între
silvicultori..........................................39
ISSN: 1583-7890
ISSN (Varianta online): 2067-1962
Indexare în baze de date:
CABI
DOAJ
Google Academic
SCIPIO ___________
Reproducerea parțială sau totală a articolelor sau ilustrațiilor poate fi făcută cu acordul redacției revistei. Este obligatorie
menționarea autorului și a sursei. Articolele publicate de Revista Pădurilor nu angajează decât responsabilitatea autorilor
lor.
Revista Pădurilor 138(3) (2023) 001-044
Revista Pădurilor
available online at: www.revistapadurilor.com
Journal edited by the "Progresul Silvic" Society
EDITORIAL BOARD
Editor in Chief:
Prof. Dr. Stelian A. Borz
Editorial Members:
Prof. Dr. loan V. Abrudan
Eng. Codruț Bîlea
Prof. Dr. Alexandru L. Curtu
Assist. Prof. Dr. Mihai Daia
Assist. Prof. Dr. Gabriel Duduman
Assist. Prof. Dr. Sergiu Horodnic
CONTENTS
Gabriel Osei Forkuo
A systematic survey on convențional and new postural assessment
methods.........................................................
Elena Camelia Mușat
Prcfessor dr. Arcadie Ciubotarii at the age cf75...................35
Dan-Marian Gurean, Pantelimon Ularu, Victor Adrian Indreica
Prcf dr. biol. Marius Danciu (1941-2023), a botanist among
foresters...............................................39
ISSN: 1583-7890
ISSN (ONLINE): 2067-1962
Indexed by:
CABI
DOAJ
Google Academic
SCIPIO
Paitial 01 full reproduction cf articles and figures may be done pending the agreement cf the journal. In case cf parțial or
full reproduction, it is mandatory to mention the authors and the source article. Articles published by Revista Pădurilor
hold solely the responsibility cftheir authors.
Revista Pădurilor 138(3) (2023) 001-044
Revista Pădurilor
available online at: www.revistapadurilor.com
A SYSTEMATIC SURVEY OF CONVENȚIONAL AND NEW
POSTURAL ASSESSMENT METHODS
Gabriel Osei Forkuoa*
aDepartment of Forest Engineering, Forest Management Planning and Terrestrial Measurements, Faculty
of Silviculture and Forest Engineering, Transilvania University of Brașov, Șirul Beethoven 1, Brașov
500123, Romania, gabriel.forkuo@unitbv.ro (G.O.F.).
HIGHLIGHTS
•	Postural assessment is crucial in
evaluating the ergonomie condition in
forest activities.
•	A range of tradițional and modern
methods are used for postural
assessment.
•	Machine Learning (ML) algorithms
and Computer Vision (CV) are
emerging techniques for postural
assessment that allow for analysis of
data from sensors and digital
images/videos.
ARTICLE INFO
Article history:
Manuscript received: 22 September 2023
Received in revised form: 25 September 2023
Accepted: 25 September 2023
Page count: 34 pages.
Article type: Review
Editor: Stelian Alexandru Borz
Keywords:
Postural assessment
Tradițional methods
Machine learning
Algorithms
Computer vision
GRAPHICAL ABSTRACT
Postural assessment method
ABSTRACT
Postural assessment is essential for evaluating the ergonomie
condition cfwork systems, especially in forest activities. Tradițional
and state-cf-the-art methods such as seif-reports, observational
techniques, motion capture systems, electromyography, and force
plates are used. Machine Learning (ML) algorithms can analyse data
from sensors to find trends or abnormalities in postare, whereas
Computer Vision (CV) extracts irformation from digital images and
videos. This systematic review identfies studies that have used ML
and CV for postural assessment, appraises their quality, and
synthesizes available evidence on the accuracy and reliability cf
dijferent techniques. The review also documents the potențial
benefits and limitations cfML and CV in postural assessment. The
review highlights the importance cf large and diverse datasets,
ethical considerations, and the interpretability cfML models. Gaps
and the needforfurther research are also idenhfied.
* Corresponding author.
E-mail address: gabriel.forkuo@unitbv.ro
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
1.	INTRODUCTION
Postural assessment is the process of observing and evaluating a person's posture, or the
alignment and orientation of the body segments in relation to each other and the environment [1],
Postural assessment is an important element in identifying the implications of working postures for
health and safety, and it can assist in the implementation of ergonomie interventions aimed at
reducing work-related injuries in forest operations [2,3], Thus, a thorough postural assessment can
help identify high-risk postures and inform the development of ergonomie interventions.
Additionally, postural assessment can provide valuable Information about the health, function, and
performance of the musculoskeletal System, as well as the potențial risk factors for injury, pain, or
discomfort [1,3]. Postural assessment is an essential aspect of diagnosing and treating
musculoskeletal and neurological disorders, as well as monitoring the progress of rehabilitation
programs [4-6],
A variety of methods, both convențional and new, are used in postural assessment [7-9],
Convențional methods include direct observation, APM, and RULA [8,10], However, new methods
include the use of wearable sensors and motion capture technologies [8,11,10], Both convențional
and new methods have their advantages and disadvantages, but they can be effective in improving
ergonomie conditions [3], For instance, direct observation is a low-cost method that can provide
detailed information on worker postures and movements [8], while motion capture technologies can
provide accurate and objective data [10], Besides, postural assessment can be performed in various
settings, such as clinical, sports, occupational, or educațional environments, and can involve
different methods and tools, such as visual inspection, palpation, goniometry, photography, video
recording, or specialized software [3,12]. However, these methods and tools have some limitations,
such as subjectivity, variability, low accuracy, high cost, or low accessibility [3,8], Tradițional
methods of postural assessment, such as manual goniometry, REBA, RULA and OWAS are
subjective and can be unreliable and time-consuming [3].
Machine learning (ML) and computer vision (CV) are two branches of artificial intelligence
(AI) that deal with the analysis and interpretation of data using algorithms that can learn from data
and make predictions or decisions [13,14], ML is a subset of AI that analyses and learns from data
[15,16], whereas CV focuses on the processing, analysis, and understanding of digital images or
videos [17,18], ML and CV have been applied to various domains and tasks, such as face recognition,
object detection, natural language processing, medical diagnosis, or self-driving cars. Recent
advancements in computing, ML and CV have also led to the development of more objective and
innovative methods for postural assessment [5,6,12] using ML and CV. ML and CV have numerous
applications in postural assessment, including the assessment of gait patterns, balance measures,
and postural deviations [5,6,12]. These techniques can be used to develop automated Systems for
postural assessment, which can provide more accurate and efficient assessments compared to
manual assessments [19,20]. For example, one common application of ML in postural assessment is
the use of sensors, such as accelerometers and gyroscopes, to monitor subjects' postures and
movements in real-time [11,21], These sensors collect large amounts of data, which can be processed
with algorithms to defect abnormalities in posture and movement patterns [11], Similarly, CV
2
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
techniques can be used to analyze video recordings of subjects' postures, which can provide highly
accurate measurements of joint angles and other postural variables [19,20],
In recent years, the application of ML and CV to postural assessment bas offered new
possibilities and advantages for the detection, measurement, classification, and correction of
postural deviations [12,19,20], Some of the benefits of using ML and CV for postural assessment are
higher objectivity and reliability, higher accuracy and precision, higher speed and efficiency, lower
cost and complexity, higher accessibility and convenience and higher scalability and flexibility
[8,19,20], Despite the growing interest and ongoing research in this area [12], few systematic reviews
have been conducted to assess the applications of ML and CV in postural assessment. After
considering the nature and scope of this study, a systematic review was considered the most
appropriate methodological approach [22,23], In particular, a systematic review provides a
comprehensive and rigorous approach to summarizing the available evidence on a specific research
question through a predefined and reproducible selection process [23], It involves a thorough and
exhaustive search of the literature, comprehensive screening and data collection, quality appraisal,
data synthesis, and reporting procedures [22,23], This approach ensures that all eligible studies
relevant to the research question are identified and evaluated in a standardized manner [24], The
outcomes of a systematic review are valid, reliable, and evidence-based, which is why it is
considered the gold standard method for evidence synthesis [23-25], On the other hand, a scoping
review is a methodology used to identify gaps in the literature and to map the existing literature of
a topic or a research field [23], While scoping reviews provide an overview of a research field, they
do not conduct quality appraisal or assess the impact of different variables on the findings [23],
Hence, they may be less useful when providing specific recommendations for clinicians and
policymakers and may not be useful for answering specific research questions in healthcare [23],
The aim of this systematic review is to provide a comprehensive overview of the current state-
of-the-art of ML and CV applications in postural assessment. Specifically, this review addresses the
following research questions: (i) What are the tradițional methods used for postural assessment? (ii)
What are the main types of ML and CV methods/techniques and algorithms used in postural
assessment? (iii) What are the data sources for training ML and CV algorithms? (iv) What are the
main types of outcomes (e.g., accuracy scores) obtained by ML and CV techniques for postural
assessment? (v) What are the main challenges and limitations of ML and CV techniques for postural
assessment? and (vi) What are the main future directions and opportunities for machine learning
and computer vision techniques for postural assessment?
This review follows a systematic methodology based on the PRISMA guidelines [24,25] to
identify, select, appraise, synthesize, and report relevant studies from various databases. The review
also provides a criticai analysis of the quality, validity, and applicability of the studies included. The
review would contribuie to the advancement of knowledge and practice in the field of postural
assessment by providing a comprehensive summary of the current evidence and highlighting the
gaps and needs for further research. The objectives of this study are to comprehensively search the
scientific literature to identify studies that have used ML and CV algorithms in postural assessment;
critically appraise the quality of the identified studios using established frameworks and tools;
synthesize the available evidence on the accuracy and reliability of different ML and CV techniques
in postural assessment; and document the potențial benefits and limitations of ML and CV
techniques in postural assessment.
3
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
2.	MATERIALS AND METHODS
The purpose of this review was to explore the existing research on machine learning and
computer vision applications in postural assessment. To achieve this, the following steps in the
methodology were implemented in line with the methods adopted in previous studies: scoping the
literature review; adopting a search strategy and identifying Information sources; selection and
screening of literature; data extraction and analysis; and synthesis and reporting [24-26], The
literature selection process, study characteristics, results and findings were reported following the
PRISMA reporting guidelines [24-26],
A protocol was developed that outlined the objectives, scope, and methodology of the review
before it was started [24], This would help to guide the review process and maintain clarity and
focus [26], Information on the research question, inclusion and exclusion criteria, and the search
strategy for identifying relevant studies were included in the protocol. Then the research questions
or objectives that guided the literature review were defined. The scope and boundaries were
identified within the field of postural assessment and specific domains and subdomains relevant to
the research question were determined. Factors such as time frame, geographical location, and
specific populations or Industries were considered to focus the literature search.
Appropriate academic databases, registers and reference lists (e.g., PubMed, Scopus,
ScienceDirect, Google scholar, Elsevier, Web of Science, IEEE Xplorc, Springer Link) were chosen,
and a comprehensive list of keywords and synonyms related to the research topic methods of
postural assessment, and ML and CV in postural assessment were developed [24,25], Citation
chaining and reviewing gray literature sources, such as reports, theses, dissertations, and conference
papers that are relevant to postural assessment methods were also taken into consideration. The
search was conducted by executing the search strategy in the selected databases and registers using
the identified keywords and search terms such as musculoskeletal disorders, OWAS, RULA, REBA,
machine learning, deep learning, computer vision, postural assessment, gait analysis and balance assessment.
The search was limited to articles published in English without any ycar range.
A total of 182 articles were selected for inclusion in this review, based on their relevance to the
topic and methodological quality. The search results were downloadcd and stored, and duplicates
were identified and removed [24-26], After identifying relevant studies, inclusion and exclusion
criteria were specified based on the scope and research questions [27], During literature selection
and screening, specific criteria for the selection of articles were developed, including publication
date, study design, relevance to the research question on postural assessment, and quality of
methodology. Titles and abstracts were initially screened to identify potentially relevant articles,
and the full texts of the selected articles were subsequently reviewcd to determine their suitability
for inclusion in the literature review on postural assessment. Inclusion criteria considered studies
that were published on postural assessment at any time, including research conducted with humans,
related to ML and CV applications in postural assessment, and were available in English. Exclusion
criteria included studies that focused on animals, studies that were not peer-reviewed, and studies
that were not relevant to ML and CV applications in postural assessment. To extract and analyze
relevant Information, a data collection framework was created [24] to systematically document key
4
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
Information from selected articles, such as study objectives, methodology, findings, limitations, and
gaps identified by the authors relevant to postural assessment. A structured framework, such as a
table or thematic matrix, was further developed to facilitate the comparison and analysis of the
extracted Information within the context of postural assessment. In analyzing the data, common
themes, patterns, and trends within the literature on ML and CV applications in postural assessment
were identified. Findings, methodologies, theories, and gaps in knowledge identified by different
authors were compared and contrasted to determine if significant gaps, limitations, or
inconsistencies existed within the existing body of knowledge [24], Discrepancies, unanswered
research questions, contradicting findings, or emerging areas with limited study were sought in
relation to postural assessment.
Figure 1. The PRISMA flow diagram of the methodological phases of the literature review for this study.
Finally, the findings of the literature review were synthesized by summarizing the main
themes, knowledge gaps, and current research trends that emerged from the literature review within
5
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
the context of postural assessment. The PRISMA guidelines were followed to ensure transparent
and comprehensive reporting of the review methodology [24], Strengths and weaknesses in the
existing literature, including any biases, limitations, or areas requiring further investigation, were
also identified [23-25], A comprehensive literature review document, annotated bibliographies, or
thematic summaries were used to present the results of the literature review in a clear and organized
manner. The existing knowledge gaps within the research field of machine learning and computer
vision in postural assessment were identified and analyzed, providing a solid foundation for further
developing research and contributing to the advancement of the scholarly discourse in the specific
domain of postural assessment. The risk of bias and quality of the studies were assessed using
appropriate tools [23], Figure 1 is a Preferred Reporting Items for Systematic Reviews and Meta-
Analyses (PRISMA) flow diagram [28,26] that illustrates the methodological phases of the literature
review on ML and CV applications in postural assessment.
3.	RESULTS AND DISCUSSION
3.1.	Convențional methods used in postural assessment
A wide variety of methodologies used in postural assessment have been identified and
grouped into three categories that have typically been utilized by past reviewers [8,29]. Studies of
[3] and [29] categorized postural evaluation methods into three groups based on their level of
accuracy and invasiveness towards the workers being assessed. In descending order of accuracy and
invasiveness [29], they are direct measurements, observational approaches, and self-reports [3,8,30-
33], Worker self-reports can serve as a valuable postural evaluation technique for gathering
Information on their exposure to work-related physical and psychological health hazards [3,29],
Various methods can be used, including worker diaries, interviews, and questionnaires. Recent
innovations can also be used, such as the self-evaluation of video films of job [34] or the use of web-
based surveys [35,36]. Employees document their postures and motions during the workday for
worker diaries. For the use of interviews, experts or supervisors conduct interviews with workers to
gather Information on physical and psychosocial factors related to postural risks. Regarding
questionnaires, workers complete questionnaires to provide information on their work tasks,
physical demands, and psychosocial factors [3,8,32]. Some examples of studies that have used self-
reports as a method of postural assessment in forestry and office work were presented in the studies
of [35-39]. The ease of use, adaptability to various work situations, and cost-effectiveness of these
strategies for surveying a large sample size are just a few of its advantages, according to [3] and [29].
Moreover, they enable the estimation of exposure for more extended periods than what can be
observed at work [3]. Despite the usefulness of self-reports in postural assessment, it has been found
that worker perceptions of exposure can be vague and unreliable, posing challenges to self-
reporting. For example, workers experiencing significant low back or neck pain have been found to
report longer or more frequent periods of physical activity compared to pain-free workers, although
the findings of such studies may vary [38], Additionally, differences in worker literacy,
comprehension, or question interpretation can impact the reliability of self-reports [39], Although
self-reports have drawbacks, [3] and [29] both point out that they can be useful in identifying
occupational groups that are more at risk. Study of [29] reports that self-reports can be utilized to
collect information regarding demographic characteristics, reported symptoms such as pain and
6
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
postural discomfort, and levels of perceived exertion. Large sample sizes are frequently needed to
guarantee that the data collected is representative of the occupational groups under study, and the
right skills are needed to properly evaluate the results [3,29], As a result, using employee self-reports
as a tool for postural evaluation can help gather Information on how exposed employees are to
physical and psychosocial problems at work. While they have limitations in terms of their reliability
and validity, they can still be used to identify occupational groups at higher risk for further analysis
using other methods. In order to acquire a more complete picture of workplace exposure to risk
factors for work-related musculoskeletal disorders (WRMSDs), it is crucial to consider the limits of
self-reports when utilizing them as a technique of postural evaluation and to combine them with
additional approaches [3,29], More research is required to increase the accuracy and dependability
of self-reports and to find the most effective applications for them in the diagnosis and avoidance
WRMSDs.
Observational methods of postural assessment are crucial in assessing the externai exposure
of workers [3,8]. There are two types of observation methods: subjective and systematic [8],
Subjective methods ask workers directly about their workload, discomfort, and stress using body
maps, rating scales, checklists, or interviews [3,8], In contrast, systematic observations involve
recording and tracking the workers' postures and movements [8], Study of [3] reported that
systematic observational methods can be categorized into simpler and advanced methods. Simpler
methods involve using pro-forma sheets to systematically record workplace exposure, which helps
prioritize intervention strategies. Simpler observational techniques include OWAS method [40],
REBA method [41], Corlett method [42], Vira method [43,44], PATH method [45], NIOSH equation
[46,47], Snook and Ciriello tables [48], KIM method [49], MAC method [50], Liberty Mutual tables
[51], RULA method [52], "Job Străin Index" method [53], PRRI [54], IBV method [55], OCRA method
[56], PLIBEL method [57], Posturegram [58], HAMA [59], QEC [31], checklists [60], FIOH Risk Factor
Checklist [61], ACGIH TLVs [62] and Upper Limb Disorder Guidance, HSG60 [63], Some methods,
such as those used by [31] and [63], exclusively evaluate the posture of various body parts.
Conversely, other methods, as discussed by [47], gather subjective data from workers as part of the
assessment of physical or psychosocial demands. For a variety of workplaces where other
observation techniques can be disruptive, these methods are practicable and affordable [3,29], They
are best suited for evaluating static or repetitive jobs, although they may vary slightly when it comes
to selecting between various exposure level groups [64], Besides, various techniques, such as those
described by [8] and [31], allow for the calculation of overall scores for combinations of exposure
factors. These scores can then be used to determine acceptable exposure limits or the order of priority
for intervention. According to [8] and [29], the data used to create these scoring systems is sparse,
particularly when it comes to figuring out how much weight to give to particular traits or
characterizing the connections between components. Study of [3] pointed out that ergonomie
approaches can be utilized to evaluate exposure to factors that increase the risk of developing
musculoskeletal problems at work. In contrast, study of [32] highlights the importance of
considering multiple aspects and recommends more complex techniques for MSD risk assessment.
Methods such as those described by [8], [31], and [47] allow for overall indices or scores to be
determined. The ratings obtained from scoring systems are essential in determining acceptable
exposure limits for workers and selecting the most criticai activities for interventions. However, as
noted by [3] and [29], these scoring systems are based on limited epidemiological data, with a
particular lack of Information on how to weigh or quantify interactions between different
7
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
parameters. It is important to note that these scoring Systems are speculative and may not provide
an accurate representation of the numerous physical exposure issues that employees face. To
prevent WRMSDs and other Health issues, it is essential to consider a variety of variables and
continuously evaluate and improve the assessment methods [3,8,29,31].
Traditionally, postural assessment in the forestry industry has been conducted using a range
of methods, strategies, and equipment [65], For example, [65] utilized four different assessment
methods, including OWAS, Revised NIOSH Lifting Equation (RNLE), OCRA, and REBA to evaluate
the index of exposure to WMSDs and assess the risks of injuries in workplace activities. In another
study, [66] systematically compared three observational methods, OWAS, RULA, and REBA, for
assessing musculoskeletal loadings and their connection to musculoskeletal disorders (MSDs). The
finding was that RULA is the most commonly used technique among the three, although it has a
significant constraint of only having two categories for leg positions. According to [67] and [68], all
pen-and-paper techniques for postural evaluation require trained experts and are time-consuming.
Therefore, attempts have been made to automate some of these techniques to perform online
evaluations. For instance, [69] and [70] employed human motion data in both simulated and real-
world settings to integrate RULA with REBA. Nevertheless, observational approaches still have
limits because they only take task dynamics, such as constant contact forces, into restricted
consideration [71], Nonetheless, tradițional methods have been shown to be feasible due to the
availability of specific software and the ease of evaluations required [65,72], Nevertheless, there is
still a need for improving the current methods for postural assessment since there has not been much
research done in this area [65], Furthermore, there are not many studies cxamining postural risks in
forest operations in terms of human-machine interaction [73],
In order to create and improve postural evaluation techniques that can efficiently identify and
reduce the hazards of WRMSDs in the forestry industry, more study is therefore needed. Therefore,
recent research has focused on developing advanced methods that can better capture the dynamic
nature of work tasks. One potențial area for improvement is the use of wearable sensors and motion
capture technology to provide more accurate and objective measurements of postural positions and
movements [73], For example, wearable sensors have been used to monitor workers' postures and
movements at the actual time during which the event occurs [74], Virtual reality simulations have
also been used to evaluate ergonomie risk factors in a controlled and repeatable environment [8,74].
These technologies can also provide workers with immediate feedback, allowing them to correct
their posture and prevent the progression of musculoskeletal problems. By incorporating ergonomie
design principles into the development of forestry Instruments and equipment, workplace
ergonomics can be improved, reducing the risk of musculoskeletal disorders [72], Postural variation
during dynamic activities is measured using sophisticated observational techniques, and data is
captured on videotape or processed using specialist software [3,8], To provide a more accurate and
thorough evaluation of ergonomie risk factors, these methods involve computerized data collecting
and analysis. Using computerized time sampling, ROTA [75] and TRAC [64] approaches, for
instance, record and evaluate activity and posture while allowing for the assessment of both
dynamic and static tasks. To calculate internai exposures during task execution, the Biomechanical
models method [76] links segments of the human body together. Other techniques make use of
video-based techniques, like the Simi Motion technique [31], Overall, these advanced observational
methods are used to provide a more objective and comprehensive assessment of postural risk factors
8
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
in the workplace, allowing for the development of effective interventions to prevent work-related
injuries and improve worker Health and safety.
Postural evaluation techniques vary in their assessment of exposure factors, with some
considering physical and psychosocial aspects in addition to postural alterations, while others focus
solely on postural changes [71], Some techniques use video-based observational methods and
specialized software to objectively record and analyze data, while others rely on subjective data
collected from employees [53,77], Advanced methods may quantify intersegmental moments and
forces using biomechanical models as well as a variety of movement dimensions, such as distance,
angular shifts, velocities, and accelerations [78], Study of [71] emphasize that when evaluating
workplace activities, it is important to consider factors beyond movement characteristics, such as
load, repetition, movement duration, and vibration. A lack of epidemiological data limits the
usefulness of the many indices and scoring Systems that have been established to determine
acceptable exposure limits for employees or to prioritize interventions [71], Although some methods
are more exhaustive, they lack precision and reliability and are subject to discrepancies between
different observers and within the same observer [71], The EAWS and CERA methods were created
in order to give a thorough ergonomie study [79,80], The Key Indicațion Method (KIM) was created
to assess manual handling tasks such as lifting, carrying, holding, pushing, and tugging, as
described by [81] and [82], The EAWS, CERA, and KIM approaches provide a thorough ergonomie
study; nevertheless, additional research is needed to determine their validity and usefulness in
practicai contexts [8], In addition, these techniques may not always capture the nuances of
individual tasks and may require customization for specific industries and work environments.
Therefore, continuous improvement and development of postural assessment methods and
techniques are necessary to ensure the prevention of MSDs resulting from work and to enhance
workers' general well-being and security. According to [3] and [29], the use of advanced
observational methods for assessing postures can be costly and requires extensive technical support
from highly trained staff. Although these methods are relatively inexpensive and can be used in
various work situations, they have limitations in terms of reliability and validity, particularly for
subjective judgement methods. As a result, it is crucial to combine several methods, including
objective measures and self-reporting, to ensure comprehensive and accurate postural assessment
in forest operations. While some methods may be more exhaustive, they can lead to inconsistencies
between different observers and within the same observer, which can be time-consuming. Recent
technological advancements have led to the development of more objective and dynamic assessment
methods, but further research is needed to establish their validity and reliability in real-world
settings [8], Despite their drawbacks, advanced observational techniques have been found to be
better suited for capturing and assessing simulated tasks than for conducting actual workplace
assessments. This is due to the fact that their utilization might be time-consuming in practice and
necessitate intensive technical support. But using these techniques at work is crucial to preventing
musculoskeletal conditions brought on by improper posture.
Direct measurement of exposure variables can be achieved by attaching sensors to the subject's
workspace [3,8,29], Data can be collected both on-site at the workplace, and during laboratory
simulations [8], Direct methods, including hand-held Instruments that measure joint mobility and
electronic goniometers that continuously record joint movement during task performance were
reported and used in the studies of [31,86-93], Specialized lightweight devices can be attached
directly across articulating joints to measure finger, wrist, and forearm angles, as well as forearm
9
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
rotati on, with corresponding systems available for computerized data analysis [3,29],
Techniques for observing one's posture are additional direct methods for evaluating physical
workload. With these techniques, certain anatomical sites on the worker are marked with
optical, acoustic, or electromagnetic markers, which are subsequently tracked by related
scanning equipment to keep track of different body components' angles of movement and
positions. AII body markers' three-dimensional coordinates can be recorded by sophisticated
computing equipment immediately. However, according to [3] and [29], these systems are more
appropriate for task simulations than industrial setting invcstigations. Studies of [3] and [8]
have pointed out that Electromyography (EMG) is another direct method for measuring muscle
tension that involves recording and analyzing myoelectrical activity using a computer. EMG
activity can be translated into muscle tension using empirical models, although the relationship
may not be linear and needs careful interpretation. By examining changes in the spectrum
properties of the myoelectric signal, EMG can also evaluate local muscle exhaustion, however
interpretation can be difficult. In order to drive simulations of musculoskeletal models for
the upper and lower extremities, EMG-based approaches have been utilized to monitor
physical exposure and cumulative compression in the low back [3,29,94]. Additionally, the
direct method involves the use of inclinometers which measure the angle of the spine or other
body segments to determine the degree of bending or twisting, and force plates which measure
the ground reaction forces during work tasks to determine the level of physical demand [3,8,29],
While direct sensor attachment to the subject can make the subject uncomfortable and change
their working habits, direct measurement techniques provide very accurate data on a variety of
exposure characteristics. These systems produce a large amount of data, however because to the
time needed for data analysis and interpretation, practitioners may find it unworkable.
Additionally, in order to ensure effective functioning, direct measurement systems need a
sizeable inițial expenditure to pay for the equipment, cover maintenance expenses, and hire
highly qualified technical employees [3], A flexible model was created by [95] to investigate
compressive străin on the human spine. The goal of this model is to reduce complexity while
preserving accuracy for more intricate body parts. Moreover, a stochastic methodology for
ergonomie evaluation was introduced by [96] by simultaneously evaluating human whole-body
kinematics and dynamics. However, further research and development needs to be done on
these techniques before they can be completely applied in scenarios connected to forests [3,8],
In order to quantify exposure factors at work, direct approaches for postural assessment entail
directly attaching sensors to the individual. These methods provide large quantities of highly
accurate data but require considerable investment and technical expertise for effective operation
[8], Recent innovative approaches aim to account for workers' internai exposure and address
the limitations of tradițional ergonomie tools and laboratory-based methods. While these
approaches show promise, they still need further development and validation before
being widely adopted in forest industry settings. Recent advancements in sensor technologies
have enabled the automation of many methods belonging to the observational and direct
measurement categories [3,8], To better understand the physical internai exposure of
humans, researchers use sensor systems to collect direct measurements on human subjects [8],
These measurements are then integrated with complex models of the human body to estimate
muscle tensions and joint loads [3,8], OpenSim is a freely available application that can create
interactive simulations of movement utilizing models of the neuromusculoskeletal system. It
incorporates a commonly used algorithm, while Anybody, another software, views the human
musculoskeletal system as a rigid-body system. Both programs consider the muscle
recruitment patterns and geometry in their models. By utilizing techniques such as inverse
10
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
kinematics and inverse dynamics, the motion of these models can be analyzed [97,98], However,
their utilization is not strictly obligatory, as ergonomics assessments can still be approximated by
incorporating them into ad-hoc models or directly retrieved from acquired data [8], Overall, the
classification/categorization of the various methods used in postural assessment in forest operations
provides a systematic and comprehensive overview of the different approaches available. This
categorization enables a better knowledge of the advantages and disadvantages of each approach,
which can aid researchers and practitioners in choosing the approach that is best suitable for their
individual needs. By using a combination of these methods, it is possible to obtain a comprehensive
assessment of postural risks in forest operations, which can ultimately lead to the development of
effective interventions to prevent MDSs and injuries resulting from work. Generally, these methods
can be used individually or in combination to assess postural risks in forest operations, and each
method has its advantages and limitations depending on the specific needs of the study or
intervention.
The selection of postural assessment methods should be based on the particular research
subject and the environment where the evaluation is taking place. Other relevant considerations
include the level of detail required, available resources and time, and the target group being
evaluated. The selected method should allow for accurate and reliable data collection, while also
being practicai and feasible for the research context. For example, simpler methods may be more
appropriate for assessing postural risks in low-risk work environments, while more advanced
techniques may be necessary for high-risk work environments with complex postural demands. In
addition, the use of multiple methods can provide a more comprehensive assessment of postural
risks and can help to validate the findings of each individual method. Overall, a multi-method
approach can help to identify postural risk factors and develop effective interventions to prevent
and manage musculoskeletal disorders. Combining several approaches makes it feasible to get a
more comprehensive picture of the postural demands of daily tasks and to pinpoint areas that can
benefit from treatments in order to lower the risk of MSDs.
3.2.	ML and CV algorithms and methods used for postural assessment
Algorithms form the core of ML and CV systems [12,19,20,99], According to [100], “irformally,
an algorithm is any well-dtfined computațional procedare that takes some value, or set cfvalues, as input and
produces some value, or set cfvalues, as output. An algorithm is thus a sequence cf computațional steps that
transform the input into the output." Consequently, various algorithms are employed for different CV
applications, including image classification, object detection, segmentation, and more
[12,19,20,99,101], According to studies of [12,19,20,99], substanțial research has been done on CV in
the areas of pose estimation, activity recognition, pose or gesture recognition, and other topics.
Postural assessment is an important component of human health monitoring and rehabilitation, as
there is a direct correlation between poor posture and various musculoskeletal disorders [102],
Recent advances in ML and DL algorithms have enabled the development of efficient and accurate
CV methods for assessing posture from visual cues, such as images and videos of individuals [12,99],
In this context, computer vision-based algorithms have been developed to obtain and assess posture
parameters in order to be used as an assistance for physical therapists in their posture assessment
and correction [99], These algorithms have been used in a variety of fields, including healthcare,
11
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
sports, and ergonomics, to analyze and improve posture in different settings [12,19,20,99,101]. They
have shown promising results in assessing for instance sitting posture [101] and detecting postural
deformity in the spine [103],
Some of the common tradițional ML algorithms used in CV for postural assessment research
include SIFT, HOG, SVM, GMM, DTW and HMM [12,104], Other tradițional ML classification
algorithms that are utilized for postural assessment using CV include RF, DT, k-NN, the Bayesian
classification algorithm, naive Bayes (NB), and linear discriminant analysis (LDA) among others.
Besides, a large number of researchers utilize both DL algorithms and narrow neural networks in
CV for postural assessment research [12], In this regard, researchers have proposed several
Convolutional Neural NetWork (CNN) algorithms and architectures, and "improved CNN methods,
such as stacked hourglass networks, multi-stage pose estimation networks, convolutional pose machines, and
high-resolution nets" [12], Furthermore, researchers have explored the use of explainable deep neural
networks, graph neural networks, transfer learning (TL), ensemble learning (EL), and other
advanced neural network applications for posture Identification [12],
Regarding the tradițional ML algorithms used in CV applications, SIFT is a feature detection
and extraction algorithm that captures all significant local features in an image [105,106], It detects
and describes local features within an image, which are invariant to scale, rotation, and affine
transformations [105,106], It has been used in posture analysis to defect key points of the humans.
These key points can be extracted using SIFT to provide feature vectors that are used to classify
posture [106], Moreover, HOG is an algorithm used for detecting objects, particularly for detecting
humans [107], HOG is a feature descriptor that represents the presence of gradients in an image
[107], It calculates the distribution of gradient orientations within an image, which is then used to
create feature vectors for classification [108], It has been used for feature extraction in identifying
human posture. HOG features can be extracted from a body part's image and trained with a classifier
to recognize different postures [108], According to [109], SVMs are commonly used classification
algorithms for image classification in computer vision. In posture recognition, SVMs have been used
to perform binary classification to determine whether a posture is correct or incorrect. Apart from
posture classification, SVMs have also been widely applied in various CV applications like object
detection and face recognition [110], Similarly, GMM is a probabilistic model that has been used for
multimodal posture modeling [111,112], Gaussian components of the model represent different
posture modes, and the model is trained with posture observations to estimate the probability of
each component [111,112], GMM can be used to identify the most probable posture mode for each
observed posture [12], DTW is a similarity measure that has been used to measure the similarity
between sequence data [113,114], DTW aligns two sequences and computes the distance between
them according to the minimum warping path. DTW has been used for measuring the similarity
between two postures and identifying any deviations between them [113,115], Furthermore, the
temporal dependencies between postures have been modeled using the Hidden Markov Model
(HMM) statistical model [116,117], The HMM consists of a set of States that represent various
postures, and transitions between the States represent the likelihood of a posture change. In order to
anticipate the sequence of postures based on the observed data, HMM has been employed in posture
recognition [116,117], The HMM algorithm is a statistical model used for time-series data that
considers uncertainty in postural data analysis to provide more accurate results [118-121], HMMs
can be used in both supervised learning (SL) and unsupervised learning (UL) settings, with the
model trained using labeled data for SL and unlabeled data for UL [119,121], HMMs have been used
12
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
in both supervised and unsupervised learning settings for postural assessment tasks, achieving high
accuracy rates in various studies [118,122-124]. HMMs are a flexible ML algorithm that can model
the temporal sequence of postural transitions and handle missing data and noisy sensor signals,
making them suitable for wearable sensor-based postural assessment [119-121], The utilization of
HMMs for physical activity classification using on-body accelerometers has potențial applications
in healthcare and sports [119], Additionally, HMMs have been applied to UL to examine temporal
patterns among posture sequences to identify categories associated to a child's level of interest, with
overall accurate results of 76.5 - 82.3% [120],
RF are ensemble learning algorithms that consist of multiple decision trees (DT), where each
tree independently classifies an input image or extracts features [125]. It is utilized for tasks such as
classification, regression, and others that generate a lot of DT during the training phase [125-128],
According to [125], RF are renowned for their dependability and capacity to handle high-
dimensional data. Studies show that numerous DT are integrated in RF to improve the model
categorization's accuracy [125,127-129]. By adding up each tree's individual decisions, the final
choice is made. In postural assessment, RF which are supervised learning algorithms have been used
to classify postures based on different features extracted from images or videos. For example, RF
have been used in the context of postural assessment to classify different postures based on various
motion capture data [5,12]. They have also been used to predict postural sway in individuals with
Parkinson's disease [5], They are particularly helpful for activities involving a lot of characteristics
or variables. In the context of forest operations, RF could be used to classify different postures based
on data collected from wearable sensors [128], They are used in CV for tasks like object detection
and image classification. Several studies report the use of RF in CV for postural assessment research
[129-131], DT are graphical models that split the data into branches based on some criteria, such as
entropy or Gini index [132,133], Each branch represents a possible outcome or a decision rule
[132,133], The leaf nodes at the end of the branches indicate the predicted class or value for the data.
Studies of [132] and [133] have noted that DT are straightforward to interpret and can handle both
categorical and numerical data. However, they are prone to overfitting and instability [132,133]. DT
have been utilized in CV for postural assessment research [12,134-136], For instance, [134] used DT
to classify different postures of workers, such as standing, sitting, bending, etc. in a manufacturing
environment. They achieved an accuracy of 94.4% on a dataset of 1200 images. Similarly, [135] used
DT to classify the posture quality of office workers based on their head, neck, and trunk angles. They
achieved an accuracy of 86.7% on a dataset of 300 images. Additionally, Singh and Garg (2016) used
DT to classify the posture of drivers based on their steering wheel grip, head position, and seat belt
usage. They achieved an accuracy of 88.9% on a dataset of 180 images. The K nearest neighbors
(KNN) algorithm, which is based on distance, assigns a new data point to the category or value of
its K nearest neighbors in the training data [17], According to [137] and [138], KNN is another
algorithm used for classification of a new data point in light of its closest neighbors' classes. The
value of K is a hyperparameter that can be chosen based on the data distribution or cross-validation
[17], KNN is simple and flexible, but it can be slow and sensitive to noise and outliers [17], Study of
[139] implemented the KNN algorithm, in addition to SVM, DT, and RF, to classify the posture of
elderly individuals in postural assessment research. Their study achieved high levels of detection
accuracy (97.34%), precision (98.50%), recall (97.33%), and FI score (97.91%). Moreover, [137] utilized
many ML algorithms, such as k-nearest neighbor, to classify different activities and movements
based on data from the activity monitor. The k-nearest neighbor algorithm demonstrated high
13
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
accuracy in classifying different activities and movements. Likewisc, [140] evaluated the
identification of work posture and musculoskeletal disorder risk levels in workers in the fireworks
industry. The authors employed machine learning algorithms, including K nearest neighbors to
classify different work postures and evaluate musculoskeletal disorder risk levels. The study found
that K nearest neighbors algorithm demonstrated high accuracy in classifying different work
postures. The Bayesian classification algorithm is another probabilistic method that makes use of
Bayes' theorem to establish the posterior probability of each class in light of the properties of the
data [141], Study of [141] asserts that given the class identification, it is presumed that the traits are
conditionally independent. It can handle uncertainty and missing data, according to [141], although
it can have problems with the zero-frequency problem and unrealistic independence assumption. In
CV, the Bayesian classification algorithm has used for postural assessment research [12], For
example, study of [142] used Bayesian classification algorithm to classify the posture of patients with
spinal cord injury based on their electromyographic signals. They achieved an accuracy of 83.3% on
a dataset of 48 signals. A special instance of the Bayesian classification algorithm known as Naive
Bayes (NB) operates under the presumption that, given the class labei, all features are independent.
By multiplying the class's prior probability and each feature's likelihood given the class, it
determines each class's posterior probability. It is fast and simple, but it can be inaccurate if the
features are correlated or have zero-frequency values. In the context of postural assessment, study
of [143] used naive Bayes to classify the posture of workers in an assembly line based on their wrist
angles and muscle activities. They achieved an accuracy of 91.7% on a dataset of 120 signals. Study
of [134] used naive Bayes to classify different postures of workers in a manufacturing environment,
such as standing, sitting, bending, etc. They achieved an accuracy of 92% on a dataset of 1200 images.
According to [144], LDA is a method for classifying data and reducing its dimensionality that
projects the data onto a lower-dimensional space, maximizing variation between classes and
minimizing variance within classes. The data must have a multivariate normal distribution, and the
classes must have equal covariance matrices, according to the LDA. Inaccuracy may result if the
assumptions are broken, although it can minimize computațional complexity and prevent
overfitting. One example of using LDA for postural assessment research is that of [145] who used
LDA to classify the posture of elderly people based on their body joint angles and center of mass.
They achieved an accuracy of 93.8% on a dataset of 240 images. Moreover, the study of [146] used
LDA to classify the posture of office workers based on their head, neck, and trunk angles. They
achieved an accuracy of 87.5% on a dataset of 300 images.
However, the study of [12] suggest that tradițional ML algorithms are not always effective in
CV for postural assessment research. They argue that these algorithms have limited ability to
represent complex and semantically rich information, which can binder accurate recognition of
features. Additionally, step-by-step recognition methods may not have robust real-time
performance, posing further challenges for postural assessment. The study of [109] has argued that
deep learning-based postural assessment approaches can improve the performance of postural
assessment systems. They explain that using deep neural networks permits combining low-level
features from the image with higher-level estimates, generating more accurate results. Compared to
tradițional ML algorithms, deep learning-based models can adapt better to new tasks and perform
faster and more accurately in target detection. Consequently, the optimization of ML algorithms in
postural assessment research mainly employs neural network models as parameter structures. The
study of [12] cite CNN, DNN, RNN, LSTM, TL, and attention models as commonly used in postural
14
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
assessment research using DL. These models are implemented as a complete learning approach,
where feature extraction and model learning are automatically performed by the algorithm without
manual operation. The study of [147] explains that by starting directly from the original input data,
a hierarchical network extracts features and learns the model. DL has achieved impressive results in
postural assessment research [12,148] and the study of [12] claims that DL-based identification
posture can swiftly compare human posture data to sample labels to create a model that can assess
posture. CNNs have demonstrated a great deal of success in postural assessment and identification,
according to [12], Convolutional layers, which extract useful features from the input images, are
followed by pooling layers and fully connected layers for classification in CNNs. CNNs have
achieved state-of-the-art performance on various CV tasks [149]. As observed by [150], CNNs have
evolved over the years, leading to the emergence of various architectures and methods for a wide
range of CV applications. As a result, various improved CNN methods, such as "stacked hourglass
networks, multi-stage pose estimation networks, convolutional pose machines, and high-resolution nets",
YOLO, among others have been proposed and used for postural [12], There are many examples of
studies using DL on various datasets [152-162], Stacked Hourglass Networks (SHN) is a popular
CNN architecture used to estimate human posture [12,163]. SHN uses a multistage architecture to
refine the prediction of human pose by regressing heatmaps of body joints. The network consists of
several stacked hourglass structures, each with a series of encoding and decoding stages [163], The
network generates confidence maps for each joint's position and size in the image. According to
research by [163], SHN performs at the cutting edge on baselines for posture estimation like MPII
and COCO. Multi-Stage Pose Estimation Networks (MSPE) is another CNN-based method for
human pose estimation [164], MSPE divides the human body into several sub-regions and estimates
the pose of each sub-region before aggregating the final position of the joints [164]. The network
consists of multiple sub-networks, each focusing on a separate sub-region of the body. MSPE has
been shown to improve the accuracy and robustness of pose estimation by providing more localized
features [164], Convolutional Pose Machines (CPM) is a CNN-based method for estimating multiple-
person poses [165], CPM processes the image using a series of convolutional layers and then applies
a learning strategy that goes through several stages to improve the estimate of each joint's position
[165], The network incorporates features from lower layers into higher layers by using skip
connections to pass Information between the stages [165], CPM has been shown to achieve modem
performance on baselines for postural assessment of several people such as MPII and COCO [165].
High-Resolution Nets (HRN) is a CNN-based architecture designed for high-resolution image
inputs, such as those encountered in human pose estimation applications [166]. HRN uses dilated
convolutions to capture fine-grained details in the input image at different scales [166]. The network
consists of a series of high-resolution feature maps that improve the accuracy of the final pose
estimate. HRN has been shown to produce more accurate results than other CNN architectures on
high-resolution pose estimation benchmarks such as the MPII dataset [166]. You Only Look Once
(YOLO), according to [167], is an object identification algorithm that detects several items in an
image in real-time and predicts their bounding boxes and class probabilities in a single pass. To
anticipate object classes and bounding box coordinates, YOLO divides the input image into a grid
and applies convolutional neural networks to each grid cell. The versions of YOLO include YOLOv7,
PP-YOLOv2, and YOLOv8. It is well-known for being quick and effective at object detection tasks,
making it appropriate for real-time uses like video surveillance. Although YOLO's high level of
generalizability makes it less prone to failure when applied to new domains or unpredictable inputs,
it may perform worse than state-of-the-art detection Systems regarding accuracy, particularly in
15
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
localizing smaller objects. The study of [168] has demonstrated the potențial applicability of YOLO
in human body posture recognition and tracking, achieving an average precision of 91.7% on a
public dataset, and demonstrating the potențial application of their framework in UAVs for
surveillance and monitoring tasks. OpenPose is an open-source library that uses DL algorithms for
multi-person 2D pose detection, tracking, and key point estimation from video feeds [19,20,169],
OpenPose features pre-trained models and can use a variety of input sources, including pre-
recorded video files as well as real-time streams from webcams and other devices. Its ability to defect
up to 135 keypoints across body, hands and feet makes it suitable for several applications like action
recognition, augmented reality, and motion capture [169]. Backpropagation (BP) is a supervised
learning algorithm which makes use of gradient descent and is commonly used for artificial neural
networks. The process of training neural networks for various applications often involves
calculating the gradient of the error function based on the weights of the network. This is frequently
used in the field of ML [12], Regression, classification, and picture recognition are just a few of the
potențial jobs here. Study of [170], for instance, classified six different human postures using the
angular Information along with SVM and BP. Moreover, the study of [171] created a posture
recognition method that applies BP, SVM, and Naive Bayes without the need for feature extraction
by using skeletal data as the classifier's input data.
Therefore, the use of advanced DL algorithms in CV applications for postural assessment has
shown tremendous promise. The accuracy of posture assessment can be increased by using these
algorithms, which can automatically extract complicated and abstract Information from photos
while using less computer power. However, they also have several drawbacks, including the need
for a sizable dataset, vulnerability to adversarial assaults, and a propensity for overfitting
parameters. In general, the application of sophisticated algorithms to postural assessment has the
potențial to increase its precision and effectiveness as well as aid in the more accurate identification
and treatment of associated diseases. However, according to [12], CNN models' intricate design can
impact postural assessment research and practice, leading to a deeper model with more parameters
that demand greater computațional power. Lightweight CNN models like Xception [172], Mobile
Net [173], Shuffle Net [174], Shuffle Net V2 [175], and Squeeze Net [176] arebecoming more popular
due to the rise of loT technology and resource-constrained platforms. Thus, the emergence of these
lightweight convolutional neural network models has allowed for more efficient and effective
postural assessment research on resource-constrained platforms [12], These models have been
shown to be effective for posture recognition and can help mitigate the issue of limited computing
resources. Future research can explore the potențial of these models in addressing the challenges of
postural assessment in various contexts.
3.3.	Challenges and advancements in ML and CV applications for postural
assessment
The use of ML and CV software and tools provides several benefits for data analysis, including
the analysis of large and complex datasets, task automation, and user-friendly interfaces [12,177].
ML and CV algorithms can be used in postural assessment to identify movement patterns, defect
deviations from normal alignment, and predict the risk of falls [12,177], However, there are some
drawbacks, including the computațional demands of ML and CV methods, the importance of data
16
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
quality and quantity, and the cost of some ML and CV tools and software [129,178], However, there
are new trends and technologies that could transform ML and CV as they continue to develop.
Deep learning, which involves training algorithms using big data sets and neural networks, is
one of these new trends [12,179], The creation of new algorithms that are more effective and efficient
in processing vast amounts of data is another trend [12,180,181], These new trends and technologies
have the potențial to have an impact on a number of Industries, including ergonomics, healthcare,
finance, and transportation [182],
According to [179], recent advancements in CV algorithms have been facilitated by DL
techniques, particularly CNNs, which have integrated various architectures to revolutionize tasks
including semantic segmentation, object detection, and picture categorization. These algorithms can
learn complex visual patterns and features, making visual analysis more accurate and robust.
However, CV algorithms still face several challenges, as noted by [154], These challenges include
occlusion, variations in lighting conditions, scalability, and processing time, especially in
applications that require immediate responses such as autonomous vehicles or surveillance systems.
Additionally, the study of [173] highlights that adapting to different lighting conditions is important
for reliable CV systems, and the efficient processing and analysis of large-scale datasets requires
efficient algorithms, distributed computing frameworks, and parallel processing techniques.
According to [167], achieving real-time performance while maintaining accuracy and
robustness is complex and requires optimization of algorithms and hardware resources. The study
of [150] notes that challenges related to data limitations and quality, generalization to new scenarios,
interpretability and transparency, and adversarial attacks are of significant concern. Furthermore,
computațional requirements pose a serious challenge to CV since DL models often require massive
amounts of computațional resources for training and inference. Ethical and societal implications of
CV technologies also arise, required tobe addressed by researchers. Moreover, the integration of CV
with other domains, such as natural language processing and robotics, will lead to more advanced
and intelligent systems that can understand and interact with the world in a more human-like
manner [18,162]. According to [181], CV algorithms can be harnessed to develop more advanced
and intelligent systems that can better understand and interact with the world.
Future research in ML and CV applications in postural assessment aims to explore and
develop new network types and architectures beyond current deep learning models, as noted by
[150]. Furthermore, researchers aim to refine and adapt ML and CV techniques for more specialized
scenarios, cross-domain collaboration between CV and other domains of ML is gaining attention.
Enhancing the interpretability of ML and CV models is crucial for better understanding their
decisions, and further research is dedicated to developing techniques that facilitate model,
visualization, and interpretability, ensuring transparency and trust in ML and CV systems. Finally,
researchers aim to address challenges related to training and deploying large-scale ML and CV
models, including efficient optimization methods, distributed computing, and hardware
acceleration.
17
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
4.	CONCLUSIONS
Postural assessment is a key component of ergonomics that can help identify high-risk
postures and inform the development of interventions. Moreover, a postural assessment can provide
valuable Information on worker postures and movements, as well as on the effectiveness of
ergonomie interventions. Both convențional and new methods can be used in postural assessment
and can be effective in improving ergonomie conditions. Therefore, it is important to continue to
develop and refine methods for postural assessment to ensure that they are effective and practicai
for use in forest operations. ML algorithms are useful for a variety of applications because they can
learn from their environment and adapt to changes. With the ability to automatically identify
postures from photos or videos, ML has emerged as a key technique for postural recognition. CV is
a rapidly evolving field that encompasses algorithms and techniques for analyzing and interpreting
visual data. By leveraging diverse data sources, advanced algorithms, and efficient data processing,
CV applications can achieve accurate and reliable results. The availability of different software
platforms provides flexibility and convenience in developing and deploying CV applications across
various environments. As ML and CV continue to advance, they hold immense potențial to
transform numerous Industries and enhance human-computer interaction. Moreover, ML and CV
algorithms have made significant advancements, particularly with the integration of DL and CNNs.
Despite challenges like occlusion, lighting conditions, scalability, and real-time processing, ML and
CV continue to evolve and find applications in various fields. The future of ML and CV holds
promise with emerging technologies and interdisciplinary integration. ML and CV algorithms are
poised to play a pivotai role in fields such as postural assessment and recognition, and healthcare.
Though these algorithms have the potențial to revolutionize industries, improve efficiency, and
enhance human experiences, the prevailing challenges still need to be addressed. The future of ML
and CV looks promising with the emergence of technologies like 3D vision and AR, as well as
collaborations with other domains.
SUPPLEMENTARY MATERIALS
Not the case.
FUNDING
This work received no externai funding.
CONFLICT OF INTEREST
The authors declare no conflict of interest.
18
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
EXTENDED ABSTRACT - REZUMAT EXTINS
Titlu în română: O sinteză a metodelor tradiționale și moderne folosite în evaluarea ergonomic-posturală
Introducere: Evaluarea posturilor adoptate în muncă este importantă pentru caracterizarea condițiilor
ergonomice de desfășurare a muncii, mai ales în sectorul forestier. în ultimul timp, pentru evaluarea posturilor de muncă
s-au folosit metode tradiționale cum ar fi evaluarea pe bază de auto-raport, prin metode observaționale, sisteme de
detectare a mișcării și electromiogrcfie. Dezvoltarea tehnologică a condus, în ultima perioadă, la apariția unor metode noi
care au la bază învățarea și viziunea computerizată. Aceste metode cferă posibilitatea analizării datelor colectate cu senzori
și/sau sub formă de imagini sau fișiere video, permițând evaluarea non-invazivă, caracterizată de costuri reduse.
Materiale și metode: Lucrarea de față utilizează metodologia PRISMA pentru a colecta și sintetiza informații
cu privire la precizia și fiabilitatea diferitelor metode utilizate în evaluarea ergonomic-posturală, incluzând aici metodele
bazate pe învățare și viziune computerizată, cu scopul de a identifica beneficiile și limitările metodelor respective. în plus,
studiul evaluează importanța mărimii seturilor de date, aspectele de natură etică și interpretabilitatea modelelor de
învățare automată, ca aspecte esențiale ale evaluării ergonomic-posturale.
Rezultate și discuții: Algoritmii specifici învățării și viziunii computerizate identifică posturile care pot cauza
probleme de natură musculoscheletală fiind utili în dezvoltarea de acțiuni preventive prin educație specializată, evaluare
ergonomică și planificare. Mărimea și diversitatea seturilor de date sunt elemente care prezintă o importanță mare în ceea
ce privește performanța modelelor utilizate. Studiul evidențiază beneficiile metodelor ce au la bază învățarea și viziunea
computerizată în identificarea de trenduri și anomalii posturale care sunt mai greu de detectat prin folosirea de metode
tradiționale. De asemenea, metodele moderne implică costuri reduse, sunt non-invazive și pot opera cu seturi mari de
date. Cu toate acestea, precizia și fiabilitatea acestor metode relativ noi sunt încă discutabile iar considerentele etice ce
caracterizează protecția datelor reprezintă o prioritate. Operarea cu astfel de metode necesită transparență și
responsabilitate în etapele de dezvoltare a modelelor.
Concluzii: Metodele are au la bază învățarea și viziunea computerizată sunt caracterizate de o serie de avantaje.
Lucrarea de față caracterizează potențialul acestor metode de a îmbunătăți precizia și fiabilitatea evaluărilor posturale.
Principalele probleme care trebuie avute în vedere atunci când se utilizează aceste metode sunt cele legate de protecția
datelor, etică și interpretabilitatea modelelor.
Cuvinte cheie: evaluare posturală, metode tradiționale, învățare și viziune computerizată, algoritmi.
REFERENCES
1.	Physiopedia, 2023: Posture. Available online at: https://www.physio-
pedia.eom/Posture#cite_note-l (Accessed May 15, 2023).
2.	EU-OSHA, 2023: Occupational safety and health in Europe: state and trends 2023. Available
online at: https://osha.europa.eu/en/publications/occupational-safety-and-health-europe-
state-and-trends-2023 (Accessed May 26, 2023).
3.	David G.C., 2005: Ergonomie methods for assessing exposure to risk factors for work-related
musculoskeletal disorders. Occupational medicine, 55(3),	190-199.
https://doi.org/10.1093/occmed/kqi082
4.	Dang Q., Yin J., Wang B., Zheng W., 2019: Deep learning based 2d human pose estimation: A
survey. Tsinghua Science and Technology, 24(6),	663-676.
https://doi.org/10.26599/TST.2018.9010100
5.	Zhang D., Wu Y., Guo M., Chen Y., 2021a: Deep learning methods for 3D human pose
estimation under different supervision paradigms: a survey. Electronics, 10(18), 2267.
19
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
6.	Zheng C., Wu W., Chen C., Yang T., Zhu S., Shen J., Kehtarnavaz N., Shah M., 2020: Deep
learning-based human pose estimation: A survey. ACM Computing Surveys.
https ://doi.or g/10.1145/3603618
7.	Joshi M., Deshpande V., 2019: A systematic review of comparative studies on ergonomie
assessment techniques. International Journal of Industrial Ergonomics, 74, 102865.
https://doi.Org/10.1016/j.ergon.2019.102865
8.	Lorenzini M., Lagomarsino M., Fortini L., Gholami S., Ajoudani A., 2023: Ergonomie human-
robot collaboration in industry: a review. Frontiers in Robotics and AI, 9, 262.
https://doi.org/10.3389/frobt.2022.813907
9.	Moreira R., Fialho R., Teles A.S., Bordalo V., Vasconcelos S.S., de Morais Gouveia G.P., Bastos
V.H., Teixeira S, 2022: A computer vision-based mobile tool for assessing human posture: A
validation study. Computer Methods and Programs in Biomedicine, 214, 106565.
https: //doi. or g/10.1016/j. empb .2021.106565
10.	Rybnikăr F., Kacerovă L, Hofejsi P., Simon M., 2022: Ergonomics evaluation using motion
capture technology - literature review. Applied Sciences, 13(1),	162.
https://doi.org/10.3390/appl3010162
11.	Liu S., Tang Z., Zhang B., Liang Y., Gu X., 2022: Ergonomie design of cab structure for wheeled
combine harvester. Engenharia Agricola, 42.	https://doi.org/10.1590/1809-4430-
Eng.Agric.v42n3e20220023/2022
12.	Jiang X., Hu Z., Wang S., Zhang Y., 2023: A survey on artificial intelligence in posture
recognition. Computer modeling in engineering & Sciences: CMES, 137(1), 35.
https://doi.org/10.32604/cmes.2023.027676
13.	Alpaydin E., 2021: Machine learning (revised and updated edition). MIT Press.
https://doi.org/10.7551/mitpress/13811.001.0001
14.	Chollet F., 2021: Deep learning with Python. Simon and Schuster. Available online at:
http://silverio.net.br/heitor/disciplinas/eeica/papers/Livros/jGhollet]-
Deep_Learning_with_Python.pdf (Accessed June 21, 2023).
15.	Geron A., 2022: Hands-on machine learning with Scikit-Learn, Keras, and TensorFlow. O'Reilly
Media, Inc. ISBN-10 1098125975
16.	Goodfellow L, Bengio Y., Courville A., 2016: Deep learning. MIT press. Available online at:
https://book.huihoo.com/deep-learning/front_matter.pdf. (Accessed June 14, 2023).
17.	IBM, 2023:	What is computer vision? Available online at:
https://www.ibm.com/topics/computer-vision#:~:text=Computer (Accessed July 1, 2023).
18.	Szeliski R., 2022: Computer vision: algorithms and applications 2nd ed. Springer Nature.
Available online at: https://szeliski.org/Book/ (Accessed June 29, 2023).
19.	Ma N., Wu Z., Cheung Y.M., Guo Y., Gao Y., Li J., Jiang B., 2022: A survey of human action
recognition and posture prediction. Tsinghua Science and Technology, 27(6), 973-1001.
https://doi.org/10.26599/TST.2021.9010068
20
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
20.	Song L., Yu G., Yuan J., Liu Z., 2021: Human pose estimation and its application to action
recognition: A survey. Journal of Visual Communication and Image Representation, 76,103055.
https://doi.Org/10.1016/j.jvcir.2021.103055
21.	Wang Z., Yang Y., Liu Z., Zheng Y., 2023: Deep neural networks in video human action
recognition:	a review. arXiv preprint arXiv:2305.15692.
https://doi.org/10.48550/arXiv.2305.15692
22.	Munn Z., Peters M.D., Stern C., Tufanaru C., McArthur A., Aromataris E., 2018: Systematic
review or scoping review? Guidance for authors when choosing between a systematic or
scoping review approach. BMC medical research methodology, 18,	1-7.
https://doi.org/10.1186/sl2874-018-0611-x
23.	Higgins J.P., Altman D.G., Gotzsche P.C., Juni P., Moher D., Oxman A.D., Savovic J., Schulz
K.F., Weeks L., Sterne J.A., 2011: The Cochrane Collaboration's tool for assessing risk of bias in
randomised trials. British Medical Journal, 343, d5928. https://doi.org/10.1136/bmj.d5928
24.	Moher D., Shamseer L., Clarke M., Ghersi D., Liberati A., Petticrew M., Shekelle P., Stewart
L.A., 2015: Preferred reporting items for systematic review and meta-analysis protocols
(PRISMA-P) 2015 statement. Systematic reviews, 4(1), 1-9. https://doi.org/10.1186/2046-4053-4-
1
25.	Page M.J., McKenzie J.E., Bossuyt P.M., Boutron L, Hoffmann T.C., Mulrow C.D., Shamseer L.,
Tetzlaff J.M., Akl E.A., Brennan S.E., Chou R., 2021: The PRISMA 2020 statement: an updated
guideline for reporting systematic reviews. International journal of surgery, 88, 105906.
https://doi.org/10.1136/bmj.n71
26.	Shamseer L., Moher D., Clarke M., Ghersi D., Liberati A., Petticrew M., Shekelle P., Stewart
L.A., 2015: Preferred reporting items for systematic review and meta-analysis protocols
(PRISMA-P) 2015: elaboration and explanation. British Medical Journal, 349, g7647.
https://doi.org/10.1136/bmj.g7647
27.	Tricco A.C., Lillie E., Zarin W., O'Brien K.K., Colquhoun H., Levac D., Moher D., Peters M.D.,
Horsley T., Weeks L., Hempel S., 2018: PRISMA extension for scoping reviews (PRISMA-ScR):
checklist and explanation. Annals of internai medicine, 169(7),	467-473.
https://doi.org/10.7326/M18-0850
28.	Liberati A., Altman D.G., Tetzlaff J., Mulrow C., Gotzsche P.C., loannidis J.P., Clarke M.,
Devereaux P.J., Kleijnen J., Moher D., 2009: The PRISMA statement for reporting systematic
reviews and meta-analyses of studies that evaluate health care interventions: explanation and
elaboration. Annals of internai medicine, 151(4), W-65. https://doi.org/10.7326/0003-4819-151-
4-200908180-00136
29.	Kolgiri S., Bansode S., Hiremath R., 2017: A literature review on work-related body part
discomfort with respect of postures and movements. IAETSD Journal for Advanced Research
in Applied Sciences, 4(5). Available online at:
https://www.researchgate.net/publication/320465867 (Accessed August 25, 2023).
30.	Gomez-Galan M., Perez-Alonso J., Callejon-Ferre Â.J., Lopez-Martinez J., 2017:
Musculoskeletal disorders: OWAS review. Industrial Health, 55(4),	314-337.
https ://doi.org/l0,2486/indhealth.2016-0191
21
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
31.	Li G., Buckle P., 1999: Current techniques for assessing physical exposure to work-related
musculoskeletal risks, with emphasis on posture-based methods. Ergonomics, 42(5), 674-695.
https://doi.org/10.1080/001401399185388
32.	Roman-Liu D., 2014: Comparison of concepts in easy-to-use methods for MSD risk assessment.
Applied Ergonomics, 45(3), 420-427. https://doi.Org/10.1016/j.apergo.2013.05.010
33.	Takala E.P., Pehkonen L, Forsman M., Hansson G.Â., Mathiassen S.E., Neumann W.P.,
Sjogaard G., Veiersted K.B., Westgaard R.H., Winkel J., 2010: Systematic evaluation of
observational methods assessing biomechanical exposures at work. Scandinavian journal of
work, environment & health, 3-24. Available online at: http://www.jstor.org/stable/40967825
(Accessed June 10, 2023).
34.	Kadefors R., Forsman M., 2000: Ergonomie evaluation of complex work: a participative
approach employing video-computer interaction, exemplified in a study of order pieking.
International Journal of Industrial Ergonomics, 25(4), 435-445. https://doi.org/10.1016/S0169-
8141(99)00042-6
35.	Dane D., Feuerstein M., Huang G.D., Dimberg L., Aii D., Lincoln A., 2002: Measurement
properties of a self-report index of ergonomie exposures for use in an office work environment.
Journal of Occupational and Environmental Medicine, 73-81. https://doi.org/10.1097/00043764-
200201000-00012
36.	Viikari-Juntura E., Rauas S., Martikainen R., Kuosma E., Riihimăki H., Takala E.P., Saarenmaa
K., 1996: Validity of self-reported physical workload in epidemiologie studies on
musculoskeletal disorders. Scandinavian journal of work, environment & health, 251-259.
https://doi.org/10.5271/sjweh.139
37.	Pope D.P., Silman A.J., Cherry N.M., Pritchard C., Macfarlane G.J., 1998: Validity of a self-
completed questionnaire measuring the physical demands of work. Scandinavian journal of
work, environment & health, 376-385. https://doi.org/10.5271/sjweh.358
38.	Balogh L, Orbaek P., Ohlsson K., Nordander C., Unge J., Winkel J., Hansson G.Â., Malmo
Shoulder/Neck Study Group, 2004: Self-assessed and directly measured occupational physical
activities—influence of musculoskeletal complaints, age and gender. Applied ergonomics,
35(1), 49-56. https://doi.Org/10.1016/j.apergo.2003.06.001
39.	Spielholz P., Bao S., Howard N., 2001: A practicai method for ergonomie and usability
evaluation of hand tools: a comparison of three random orbital sander configurations. Applied
occupational and environmental hygiene, 16(11),	1043-1048.
https://doi.org/10.1080/104732201753214143
40.	Karhu O., Kansi P., Kuorinka L, 1977: Correcting working postures in industry: A practicai
method for analysis. Applied Ergonomics, 8(4), 199-201. https://doi.org/10.1016/0003-
6870(77)90164-8
41.	Hignett S., McAtamney L., 2000: Rapid entire body assessment (REBA). Applied ergonomics,
31(2), 201-205. http://dx.doi.org/10.1016/50003-6870(99)00039-3
42.	Corlett E.N., Bishop R.P., Madely S.J., Maneniee L, 1979: Posture targeting: a technique for
recording working postures. Ergonomics, 22,	357-366.
https://doi.org/10.1080/00140137908924619
22
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
43.	Kilbom A., Persson J., Jonsson B., 1986: Risk factors for work-related disorders of the neck and
shoulder-with special emphasis on working postures and movements. The ergonomics of
working postures, 1, 44-53. CRC Press, 11, eBook ISBN 9780429205620. Available online at:
https://www.taylorfrancis.com/chapters/edit/10.1201/bl2565-5/risk-factors-work-related-
disorders-neck-shoulder%E2%80%94with-s (Accessed June 2, 2023).
44.	INSHT, 1998: Guia tecnica para la evaluacion y preven-cion de los riesgos relativos a la
manipulacion manual de cargas. Guias tecnicas. Madrid, Espana. Available online at:
https://cpage.mpr.gob.es/producto/guia-tecnica-para-la-evaluacion-y-prevencion-de-los-
riesgos-relativos-a-la-manipulacion-manual-de-cargas-4/ (Accessed June 3, 2023).
45.	Buchholz B., Paquet V., Punnett L., Lee D., Moir S., 1996: PATH: A work sampling-based
approach to ergonomie job analysis for construction and other non-repetitive work. Applied
Ergonomics, 27(3), 177-187. https://doi.org/10.1016/0003-6870(95)00078-X
46.	NIOSH, 1981: Work practices guide for manual lifting. 1-198. NIOSH Technical Report No. 81-
122, Cincinnati, Ohio, USA. Available online at: https://www.cdc.gov/niosh/docs/81-
122/default.html (Accessed August 12, 2023).
47.	Waters T.R., Putz-Anderson V., Garg A., Fine L.J., 1993: Revised NIOSH equation for the design
and evaluation of manual lifting tasks. Ergonomics, 36(7),	749-776.
https://doi.org/10.1080/00140139308967940
48.	Snook S.H., Ciriello V.M., 1991: The design of manual handling tasks: revised tables of
maximum acceptable weights and forces. Ergonomics, 34(9),	1197-1213.
http://dx.doi.org/10.1080/00140139108964855
49.	Jiirgens W.W., Mohr D., Pangert R., Pernack E., Schultz K., Steinberg U., 2001:
Handlungsanleitung zur Beurteilung der Arbeitsbedingungen beim Heben und Tragen von
Las-ten. LAȘI Veroffentlichung 9. Hrsg. Lănderausschuss fur Arbeitsschutz und
Sicherheitstechnik. 4. Uberarbeitete Auflage.
50.	Monnington S.C., Quarrie C.J., Pinder A.D., Morris L.A., 2003: Development of manual
handling assessment charts (mac) for health and safety inspectors. In: McCabe T, ed.
Contemporary Ergonomics. London: Taylor & Francis; 3-8, ISBN 0-415-30994-8. Available
online at: https://www.taylorfrancis.com/books/9780429071751/chapters/10.1201/bl2800-2
(Accessed June 8, 2023).
51.	Mutual L., 2007: Liberty Mutual Manual Materials Handling Tables. Retrieved September 21,
2007, from: http://libertymmhtables. libertymutual. com/CM_LM TablesWeb. Liberty-Mutual.
Manual Materials Handling Tables 2011. Available online at:
http://libertymmhtables.libertymutual.com/CM_LMTablesWeb/pdf/LibertyMutualTables.pdf
(Accessed June 1, 2023).
52.	McAtamney L., Corlett E.N., 1993: RULA: a survey method for the investigation of work-
related upper limb disorders. Applied Ergonomics, 24(2), 91-99. https://doi.org/10.1016/0003-
6870(93)90080-S
53.	Steven Moore J., Garg A., 1995: The străin index: a proposed method to analyze jobs for risk of
distal upper extremity disorders. American Industrial Hygiene Association Journal, 56(5), 443-
458. http://dx.doi.org/10.1080/15428119591016863
23
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
54.	James C.P., Harburn K.L., Kramer J.F., 1997: Cumulative trauma disorders in the upper
extremities: reliability of the postural and repetitive risk-factors index. Archives of physical
medicine and rehabilitation, 78(8), 860-866. http://dx.doi.org/10.1016/S0003-9993(97)90201-X
55.	Garcia C., Chirivella C., Page A., Moraga R., Jorquera J., 1997: Metodo ERGOIBV. Evaluation
de riesgos laborales asociados a la carga fisica Institute de Biomecanica de Valencia. Available
Online at: https ://www .prevencionintegral. com/documentacion/publicaciones/metodo-ergo-
ibv-evaluacion-riesgos-laborales-asociados-carga-fisica (Accessed June 12, 2023).
56.	Colombini D., 1998: An observational method for classifying exposure to repetitive movements
of the upper limbs. Ergonomics, 41(9), 1261-1289. http://dx.doi.org/10.1080/001401398186306
57.	Kemmlert K., 1995: A method assigned for the Identification of ergonomie hazards - PLIBEL.
Applied Ergonomics, 26(3), 199-211. http://dx.doi.org/10.1016/0003-6870(95)00022-5
58.	Priel V.Z., 1974: A numerical definition of posture. Human Factors, 16(6), 576-584.
https://doi.org/10.1177/001872087401600602
59.	Christmansson M., 1994: The HAMA-method: a new method for analysis of upper limb
movements and risk for work-related musculoskeletal disorders. In Proceedings of the 12th
Triennial Congress of the International Ergonomics Association/Human Factors Association of
Canada, August, Toronto (pp. 173-175).
60.	Keyserling W.M., Brouwer M., Silverstein B.A., 1992: A checklist for evaluating ergonomie risk
factors resulting from awkward postures of the legs, trunk and neck. International Journal of
Industrial Ergonomics, 9(4), 283-301. https://doi.org/10.1016/0169-8141(92)90062-5
61.	Ketola R., Toivonen I.V.R., Viikari-Juntura E., 2001: Interobserver repeatability and validity of
an observation method to assess physical loads imposed on the upper extremities. Ergonomics,
44(2), 119-131. https://doi.org/10.1080/00140130118669
62.	Hygienists A., 2002: Threshold limit values for Chemical substances and physical agents,
Cincinnati, Ohio. Documentation of the Threshold Limit Values & Biological Exposure Indices,
TLVs	and	BEIs.	Available	online	at:
https://hero.epa.gov/hero/index.cfm/reference/details/reference_id/192015 (Accessed June 3,
2023).
63.	Health and Safety Executive, 2002: Upper limb disorders in the workplace HSG60. Suffolk: HSE
Books. Available online at: https://www.hse.gov.uk/pubns/books/hsg60.htm (Accessed May
28, 2023).
64.	van der Beek A.J., van Gaalen L.C., Frings-Dresen M.H., 1992: Working postures and activities
of lorry drivers: a reliability study of on-site observation and recording on a pocket computer.
Applied Ergonomics, 23(5), 331-336. https://doi.org/10.1016/0003-6870(92)90294-6
65.	Gallo R., Mazzetto F., 2013: Ergonomie analysis for the assessment of the risk of work-related
musculoskeletal disorder in forestry operations. Journal of Agricultural Engineering, 44(s2),
e!47. https://doi.org/10.4081/jae.2013.389
66.	Kee D., 2022: Systematic comparison of OWAS, RULA, and REBA based on a literature review.
International Journal of Environmental Research and Public Health, 19(1), 595.
https://doi.org/10.3390/ijerphl9010595
24
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
67.	Bortolini M., Gamberi M., Pilati F., Regattieri A., 2018: Automatic assessment of the ergonomie
risk for manual manufacturing and assembly activities through optical motion capture
technology. Procedia CIRP, 72, 81-86. https://doi.Org/10.1016/j.procir.2018.03.198
68.	Malaise A., Maurice P., Colas F., Ivaldi S., 2019: Activity recognition for ergonomics assessment
of industrial tasks with automatic feature selection. IEEE Robotics and Automation Letters,
4(2), 1132-1139. https://doi.org/10.1109/LRA.2019.2894389
69.	Busch B., Maeda G., Mollard Y., Demangeat M., Lopes M., 2017: Postural optimization for an
ergonomie human-robot interaction. In 2017 IEEE/RSJ International Conference on Intelligent
Robots and Systems (IROS), 2778-2785. IEEE. https://doi.org/10.1109/IROS.2017.8206107
70.	Van de Perre G., El Makrini L, Van Acker B., Saldien J., Vergara C., Pintelon L., Chemweno P.,
Weuts R., Moons K., Dewil R., Cattrysse D., 2018: Improving productivity and worker
conditions in assembly: part 1: a collaborative architecture and task allocation. In IEEE/RSJ
International Conference on Intelligent Robots and Systems (IROS). Available online at:
http://lidl.handle.net/1854/LU-8589348 (Accessed May 27, 2023).
71.	Chaffin D.B., Andersson G.B., Martin B.J., 2006: Occupational biomechanics. John Wiley &
sons.	Available	online	at:	https://www.wiley.com/en-
us/Occupational+Biomechanics%2C+4th+Edition-p-9780471723431 (Accessed June 2, 2023).
72.	Justavino F.C., Ramirez R.J., Perez N.M., Borz S.A., 2015: The use of OWAS in forest operations
postural assessment: advantages and limitations. Bulletin of the Transilvania University of
Brașov. Series II: Forestry, Wood Industry, Agricultural Food Engineering, 8(57)2, 7-16.
73.	Micheletti Cremasco M., Giustetto A., Caffaro F., Colantoni A., Cavallo E., Grigolato S., 2019:
Risk assessment for musculoskeletal disorders in forestry: A comparison between RULA and
REBA in the manual feeding of a wood-chipper. International journal of environmental
research and public health, 16(5), 793. https://doi.org/10.3390/ijerphl6050793
74.	Dianat L, Molenbroek J., Castellucci H.I., 2018: A review of the methodology and applications
of anthropometry in ergonomics and product design. Ergonomics: an internațional journal of
research and practice in human factors and ergonomics, 61(12), 1696-1720.
https://doi.org/10.1080/00140139.2018.1502817
75.	Ridd J. E., Nicholson A.S., Montan A.J., 1989: A portable microcomputer based System for "on
site' activity and posture recording. Contemporary Ergonomics, 89, 366-71.
76.	Chaffin D., Andersson G., Martin B., 1999: Occupational biomechanics, 3rd edn. New York:
John Wiley. Available online at: https://www.amazon.com/Occupational-Biomechanics-
Chaffin-Andersson-Hardcover/dp/B011DC0X6O (Accessed June 13, 2023).
77.	Garg A., Moore J.S., Kapellusch J.M., 2017: The Revised Străin Index: an improved upper
extremity exposure assessment model. Ergonomics, 60(7),	912-922.
https://doi.org/10.1080/00140139.2016.1237678
78.	Punnett L., 2014: Musculoskeletal disorders and occupational exposures: How should we judge
the evidence concerning the causal association? Scandinavian journal of public health,
42(13_suppl), 49-58. https://doi.org/10.1177/1403494813517324
25
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
79.	Schaub K., Caragnano G., Britzke B., Bruder R., 2013: The European assembly worksheet.
Theoretical Issues in Ergonomics Science, 14(6),	616-639.
https://doi.org/10.1080/1463922X.2012.678283
80.	Szabd G., Dobd M., 2018: CERA, an integrated tool for ergonomie risk assessment. DEStech
Transactions on Social Science Education and Human Science
https://doi.org/10.12783/dtssehs/ise2018/33661
81.	Klussmann A., Liebers F., Gebhardt H., Rieger M.A., Latza U., Steinberg U., 2017: Risk
assessment of manual handling operations at work with the key indicator method (KIM-MHO)
- Determination of criterion validity regarding the prevalence of musculoskeletal symptoms
and clinical conditions within a cross-sectional study. BMC musculoskeletal disorders, 18(1),
1-13. https://doi.org/10.1186/sl2891-017-1542-0
82.	Steinberg U., 2012: New tools in Germany: development and appliance of the first two KIM
("lifting, holding and carrying" and "pulling and pushing") and practicai use of these methods.
Work, 41(Supplement 1), 3990-3996. https://doi.org/10.3233/WOR-2012-0698-3990.
83.	Armstrong T.J., 1986: Upper-extremity posture: definition, measurement and control. The
ergonomics of working postures, 59-73. CRC Press. eBook ISBN9780429205620
84.	Wiktorin C., Mortimer M., Ekenvall L., Kilbom Â., Hjelm E.W., 1995: HARBO, a simple
computer-aided observation method for recording work postures. Scandinavian Journal of
work, environment & health, 21, 440-449.
85.	Fransson-Hall C., Gloria R., Kilbom Â., Winkel J., Karlqvist L., Wiktorin C., 1995: A portable
ergonomie observation method (PEO) for computerized on-line recording of postures and
manual handling. Applied Ergonomics, 26(2), 93-100. https://doi.org/10.1016/0003-
6870(95)00003-U
86.	Biometrics L., 1997: Goniometer and torsiometer operating manual. Gwent, UK: Blackwood
Available	online	at:
https://scholar.google.com/scholar?cluster=12967852851073641382&hl=en&inst=569367360547
434339&oi=scholarr (Accessed August 17, 2023).
87.	Radwin R.G., Lin M.L., 1993: An analytical method for characterizing repetitive motion and
postural stress using spectral analysis. Ergonomics, 36(4),	379-389.
https://doi.org/10.1080/00140139308967895
88.	Bernmark E., Wiktorin C., 2002: A triaxial accelerometer for measuring arm movements.
Applied ergonomics, 33(6), 541-547. https://doi.org/10.1016/S0003-6870(02)00072-8
89.	Hasson G.-A., Asterland P., Holmer N.-G., Skerfving S., 2001: Validity, reliability and
applications of an inclinometer based on accelerometers. In: Hagberg M, Knave B, Lillienberg
L, Westberg H, eds. X2001 —Exposure Assessment in Epidemiology and Practice. Stockholm,
Sweden: National Institute for Working Life S-112 79, 405-407.
90.	Johnson P.W., Hagberg M., Hjelm E.W., Rempel D., 2000: Measuring and characterizing force
exposures during computer mouse use. Scandinavian Journal of Work, Environment & Health,
26(5), 398-405. https://doi.org/10.5271/sjweh.560
26
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
91.	Marras W.S., Fathallah F.A., Miller R.J., Davis S.W., Mirka G.A., 1992: Accuracy of a three-
dimensional lumbar motion monitor for recording dynamic trunk motion characteristics.
International Journal of Industrial Ergonomics, 9(1), 75-87. https://doi.org/10.1016/0169-
8141(92)90078-E
92.	Wells R., Norman R., Neumann P., Andrews D., Frank J., Shannon H., Kerr M., 1997.
Assessment of physical workload in epidemiologie studies: common measurement metrics for
exposure assessment. Ergonomics, 40(1), 51-61. https://doi.org/10.1080/001401397188369
93.	Freivalds A., Kong Y., You H., Park S., 2000: A comprehensive risk assessment model for work-
related musculoskeletal disorders of the upper extremities. In Proceedings of the human factors
and ergonomics society annual meeting (Voi. 44, No. 31, pp. 5-728). Sage CA: Los Angeles, CA:
SAGE Publications. https://doi.org/10.1177/154193120004403129
94.	Mientjes M.I., Norman R.W., Wells R.P., McGill S.M., 1999: Assessment of an EMG-based
method for continuous estimates of low back compression during asymmetrical occupational
tasks. Ergonomics, 42(6), 868-879. https://doi.org/10.1080/001401399185342
95.	Ventura L., Lorenzini M., Kim W., Ajoudani A., 2021: A flexible robotics-inspired
computațional model of compressive loading on the human spine. IEEE Robotics and
Automation Letters, 6(4), 8229-8236. https://doi.org/10.1109/LRA.2021.3100936
96.	Latella C., Traversare S., Ferigo D., Tirupachuri Y., Rapetti L., Andrade Chavez F.J., Nori F.,
Pucci D., 2019: Simultaneous floating-base estimation of human kinematics and joint torques.
Sensors, 19(12), 2794. https://doi.org/10.3390/sl9122794
97.	Damsgaard M., Rasmussen J., Christensen S.T., Surma E., De Zee M., 2006: Analysis of
musculoskeletal systems in the AnyBody Modeling System. Simulation Modelling Practice and
Theory, 14(8), 1100-1111. https://doi.Org/10.1016/j.simpat.2006.09.001
98.	Delp S.L., Anderson F.C., Arnold A.S., Loan P., Habib A., John C.T., Guendelman E., Thelen
D.G., 2007: OpenSim: open-source software to create and analyze dynamic simulations of
movement. IEEE transactions on biomedical engineering, 54(11),	1940-1950.
https://doi.org/10.1109/TBME.2007.901024
99.	Debnath B., O'Brien M., Yamaguchi M., Behera A., 2022: A review of computer vision-based
approaches for physical rehabilitation and assessment. Multimedia Systems.
https://doi.org/10.1007/s00530-021-00815-4
100.	Cormen T.H., Leiserson C.E., Rivest R.L., Stein C., 2022: Introduction to algorithms.
Fourth Edition. MIT Press. ISBN 9780262046305. Available online at:
http://mitpress.mit.edu/algorithms/ (Accessed June 21, 2023).
101.	Mallare J.C.T., Pineda D.F.G., Trinidad G.M., Serafica R.D., Villanueva J.B.K., Cruz
A.R.D., Vicerra R.R.P., Serrano K.K.D., Roxas E.A., 2017: Sitting posture assessment using
computer vision. In 2017IEEE 9th International Conference on Humanoid, Nanotechnology,
Information Technology, Communication and Control, Environment and Management
(HNICEM), 1-5. IEEE. https://doi.org/10.1109/HNlCEM.2017.8269473
102.	Bystrdm P., Hanse J.J., Kjellberg A., 2004: Appraised psychological workload,
musculoskeletal symptoms, and the mediating effect of fatigue: a structural equation modeling
27
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
approach. Scandinavian journal of psychology, 45(4), 331-341. https://doi.0rg/lO.llll/j.1467-
9450.2004.00413.x
103.	Kim K.H., Sohn M.J., Park, C.G., 2022: Conformity assessment of a computer vision-based
posture analysis system for the screening of postural deformation. BMC Musculoskeletal
Disorders, 23(1), 799. https://doi.org/10.1186/sl2891-022-05742-7
104.	Ding W., Hu B., Liu H., Wang X., Huang X., 2020: Human posture recognition based on
multiple features and rule learning. International Journal of Machine Learning and
Cybernetics, 11, 2529-2540. https://doi.org/10.1007/sl3042-020-01138-y
105.	Giveki D., Soltanshahi M.A., Montazer G.A., 2017: A new image feature descriptor for
content based image retrieval using scale invariant feature transform and local derivative
pattern. Optik, 131, 242-254. https://doi.Org/10.1016/j.ijleo.2016.ll.046
106.	Lowe D.G., 2004: Distinctive image features from scale-invariant keypoints. International
journal of computer vision, 60, 91-110. https://doi.Org/10.1023/B:VISI.0000029664.99615.94
107.	Zhou J., Lin K.Y., Li H., Zheng W.S., 2021: Graph-based high-order relation modeling for
long-term action recognition. In Proceedings of the IEEE/CVF Conference on Computer Vision
and Pattern Recognition, 8984-8993. https://doi.org/10.1109/CVPR46437.2021.00887
108.	Vashisth S., Saurav, S., 2018: Histogram of Oriented Gradients based reduced feature for
traffic sign recognition. In 2018 International Conference on Advances in Computing,
Communications and Informatics (ICACCI), 2206-2212. IEEE.
https://doi.org/10.1109/ICACCI.2018.8554624
109.	Jiang X., Satapathy S.C., Yang L., Wang S.H., Zhang Y.D., 2020: A survey on artificial
intelligence in Chinese sign language recognition. Arabian Journal for Science and
Engineering, 45, 9859-9894. https://doi.org/10.1007/sl3369-020-04758-2
110.	Cortes C., Vapnik V., 1995: Support-Vector Networks. Machine Learning, 20, 273-297.
https://doi.org/10.1007/BF00994018
111.	Chen Q., Sang L., 2018: Face-mask recognition for fraud prevention using Gaussian
mixture model. Journal of Visual Communication and Image Representation, 55, 795-801.
https://doi.Org/10.1016/j.jvcir.2018.08.016
112.	Lee S., Rajan S., Jeon G., Chang J.H., Dajani H.R., Groza V.Z., 2017: Oscillometric blood
pressure estimation by combining nonparametric bootstrap with Gaussian mixture model.
Computers in biology and medicine, 85,	112-124.
https://doi.Org/10.1016/j.compbiomed.2015.ll.008
113.	Ghersi L, Ferrando M.H., Fliger C.G., Arenas C.F.C., Molina D.J.E., Miralles, M.T., 2020:
Gait-cycle segmentation method based on lower-trunk acceleration signals and dynamic time
warping. Medical Engineering & Physics, 82,	70-77.
https://doi.Org/10.1016/j.medengphy.2020.06.001
114.	Ramezanpanah Z., Mallem M., Davesne, F., 2020: Human action recognition using laban
movement analysis and dynamic time warping. Procedia Computer Science, 176, 390-399.
https://doi.Org/10.1016/j.procs.2020.08.040
28
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
115.	Oszust M., Krupski J., 2021: Isolated sign language recognition with depth cameras.
Procedia Computer Science, 192, 2085-2094. https://doi.Org/10.1016/j.procs.2021.08.216
116.	Huang C.L., Wu M.S., Jeng S.H., 2000: Gesture recognition using the multi-PDM method
and hidden Markov model. Image and Vision Computing, 18(11), 865-879.
https://doi.org/10.1016/S0262-8856(99)00042-6
117.	Sănchez V.G., Lysaker O.M., Skeie N.O., 2020: Human behaviour modelling for welfare
technology using hidden Markov models. Pattern Recognition Letters, 137, 71-79.
https://doi.Org/10.1016/j.patrec.2019.09.022
118.	Eom H., Choi B., Noh J., 2014: Data-driven reconstruction of human locomotion using a
single smartphone. In Computer Graphics Forum, 33(7),	11-19.
https ://doi.or g/10.1111/cgf. 12469
119.	Mannini A., Sabatini A.M., 2010: Machine learning methods for classifying human
physical activity from on-body accelerometers. Sensors, 10(2),	1154-1175.
https://doi.org/10.3390/sl00201154
120.	Moța S., Picard R.W., 2003: Automated posture analysis for detecting learner's interest
level. In 2003 Conference on computer vision and pattern recognition workshop, 5,49-49. IEEE.
https://doi.org/10.1109/CVPRW.2003.10047
121.	Rabiner L.R., 1989: A tutorial on hidden Markov models and selected applications in
speech recognition. Proc. IEEE. 77(2), 257-286. https://doi.Org/10.1109/5.18626
122.	Kale A., Rajagopalan A.N., Cuntoor N., Kruger V., 2002: Gait-based recognition of
humans using continuous HMMs. In Proceedings of Fifth IEEE International Conference on
Automatic Face Gesture Recognition, 336-341. IEEE.
https://doi.org/10.1109/AFGR.2002.1004176
123.	Panahandeh G., Mohammadiha N., Leijon A., Hăndel P., 2013: Continuous hidden
Markov model for pedestrian activity classification and gait analysis. IEEE Transactions on
Instrumentation and Measurement, 62(5), 1073-1083. https://doi.org/10.1109/TIM.2012.2236792
124.	Taborri J., Rossi S., Palermo E., Patane F., Cappa P., 2014: A novei HMM distributed
classifier for the detection of gait phases by means of a wearable inerțial sensor network.
Sensors, 14(9), 16212-16234. https://doi.org/10.3390/sl40916212
125.	Breiman L.,	2001: Random forests. Machine learning, 45,	5-32.
https://doi.Org/10.1023/A:1010933404324
126.	Agnihotri N., 2023: Classification of machine learning algorithms. Available online at:
https://www.engineersgarage.com/machine-learning-algorithms-classification/ (Accessed 17
June 2023).
127.	Breiman L.,	2017: Classification and regression trees. Routledge.
https://doi.org/10.1201/9781315139470
128.	Paul A., Mukherjee D.P., Das P., Gangopadhyay A., Chintha A.R., Kundu S., 2018:
Improved random forest for classification. IEEE Transactions on Image Processing, 27(8), 4012-
4024. https://doi.org/10.1109/TIP.2018.2834830
29
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
129.	Imbeault-Nepton T., Maitre J., Bouchard K., Gaboury S., 2022: Filtering data bins of UWB
radars for activity recognition with random forest. Procedia Computer Science, 201, 48-55.
https://doi.Org/10.1016/j.procs.2022.03.009
130.	Nunes U.M., Faria D.R., Peixoto, P., 2017: A human activity recognition framework using
max-min features and key poses with differential evolution random forests classifier. Pattern
Recognition Letters, 99, 21-31. https://doi.Org/10.1016/j.patrec.2017.05.004
131.	Subedi S., Pradhananga N., 2021: Sensor-based computațional approach to preventing
back injuries in construction workers. Automation in Construction, 131, 103920.
https://doi.Org/10.1016/j.autcon.2021.103920
132.	Breiman L., Friedman J., Stone C.J., Olshen R.A., 1984: Classification and regression trees.
CRC press. ISBN 9780412048418
133.	Kumar G.S., 2020: Decision trees: a step-by-step approach to building DTs. Available
online at: https://towardsdatascience.com/decision-trees-a-step-by-step-approach-to-building-
dts-58f8a3e82596 (Accessed 8 July 2023).
134.	Haider D., Yang X., Abbasi Q.H., 2019: Post-surgical fall detection by exploiting the 5 G
C-Band technology for eHealth paradigm. Applied Soft Computing, 81, 105537.
https://doi.Org/10.1016/j.asoc.2019.105537
135.	Pajak G., Krutz P., Patalas-Maliszewska J., Rehm M., Pajak I., Dix M., 2022: An approach
to sport activities recognition based on an inerțial sensor and deep learning. Sensors and
Actuators A: Physical, 345, 113773. https://doi.Org/10.1016/j.sna.2022.113773
136.	Singh R., Garg D., 2016: Hybrid machine learning algorithm for human activity
recognition using decision tree and partide swarm optimization. International Journal of
Engineering Science and Computing, 6(7),	8379-8389. Available online at:
https://www.semanticscholar.org/paper/Hybrid-Machine-Learning-Algorithm-for-Human-
Using-Singh-Garg/20eeda7b817d7f922138f8d85cb7ff5550f07996 (Accessed 24 August 2023).
137.	Bussmann H.B., Reuvekamp P.J., Veltink P.H., Martens W.L., Stăm H.J., 1998: Validity
and reliability of measurements obtained with an "activity monitor" in people with and
without a transtibial amputation. Physical therapy, 78(9),	989-998.
https://doi.Org/10.1093/ptj/78.9.989
138.	Yang C.C., Hsu, Y.L., 2010: A review of accelerometry-based wearable motion detectors
for physical activity monitoring. Sensors, 10(8), 7772-7788. https://doi.org/10.3390/sl00807772
139.	Han K., Yang Q., Huang, Z., 2020: A two-stage fall recognition algorithm based on human
posture features. Sensors, 20(23), 6966. https://doi.org/10.3390/s20236966
140.	Indumathi N., Ramalakshmi R., 2021: An evaluation of work posture and
musculoskeletal disorder risk level Identification for the fireworks industry worker's. In 2021
9th International Conference on Reliability, Infocom Technologies and Optimization (Trends
and Future Directions) (ICRITO), 1-5. IEEE. https://doi.org/10.1109/ICRIT051393.2021.9596532
141.	Theodoridis S., 2015; Machine learning: a Bayesian and optimization perspective.
Academic press. ISBN 978-0-12-801522-3
30
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
142.	Dimitrijevic M., Lepetit V., Fua, P., 2006: Human body pose detection using Bayesian
spatio-temporal templates. Computer vision and image understanding, 104(2-3), 127-139.
https://doi.Org/10.1016/j.cviu.2006.07.007
143.	Balaji E., Brindha D., Elumalai V.K., Umesh K., 2021: Data-driven gait analysis for
diagnosis and severity rating of Parkinson's disease. Medical Engineering & Physics, 91, 54-64.
https://doi.Org/10.1016/j.medengphy.2021.03.005
144.	Xanthopoulos P., Pardalos P.M., Trafalis T.B., 2013: Linear discriminant analysis. Robust
data mining, 27-33. https://doi.org/10.1007/978-l-4419-9878-l_4
145.	losifidis A., Tefas A., Nikolaidis N., Pitas L, 2012: Multi-view human movement
recognition based on fuzzy distances and linear discriminant analysis. Computer Vision and
Image Understanding, 116(3), 347-360. https://doi.Org/10.1016/j.cviu.2011.08.008
146.	Lee M., Roan M., Smith B., Lockhart T.E., 2009: Gait analysis to classify externai load
conditions using linear discriminant analysis. Human movement Science, 28(2), 226-235.
https://doi.Org/10.1016/j.humov.2008.10.008
147.	Qiu S., Zhao H., Jiang N., Wang Z., Liu L., An Y., Zhao H., Miao X., Liu R., Fortino G.,
2022: Multi-sensor information fusion based on machine learning for real applications in
human activity recognition: State-of-the-art and research challenges. Information Fusion, 80,
241-265. https://doi.Org/10.1016/j.inffus.2021.ll.006
148.	Zhao J., Obonyo E., 2021: Applying incremental deep neural networks-based posture
recognition model for ergonomics risk assessment in construction. Advanced Engineering
Informatics, 50, 101374. https://doi.Org/10.1016/j.aei.2021.101374
149.	LeCun Y., Bottou L., Bengio Y., Haffner P., 1998: Gradient-based learning applied to
document recognition. Proceedings of the IEEE, 86(11),	2278-2324.
https://doi.Org/10.1109/5.726791
150.	Chai J., Zeng H., Li A., Ngai E.W., 2021: Deep learning in computer vision: A criticai
review of emerging techniques and application scenarios. Machine Learning with
Applications, 6, 100134. https://doi.Org/10.1016/j.mlwa.2021.100134
151.	Harris E.J., Khoo I.H., Demircan E., 2022: A survey of human gait-based artificial
intelligence applications. Frontiers in Robotics and AI, 8,	749274.
https://doi.org/10.3389/frobt.2021.749274
152.	Yang W., Li S., Ouyang W., Li H., Wang X., 2017: Learning feature pyramids for human
pose estimation. In proceedings of the IEEE internațional conference on computer vision, 1281-
1290. https://doi.org/10.48550/arXiv.1708.01101
153.	Wang J., Ma Y., Zhang L., Gao R.X., Wu D., 2018: Deep learning for smart manufacturing:
methods and applications. Journal of manufacturing systems, 48,	144-156.
https://doi.Org/10.1016/J.JMSY.2018.01.003
154.	Russakovsky O., Deng J., Su H., Krause J., Satheesh S., Ma S., Huang Z., Karpathy A.,
Khosla A., Bernstein M., Berg A.C., Fei-Fei L., 2015: ImageNet large scale visual recognition
challenge. International Journal of Computer Vision, 115(3),	211-252.
https://doi.org/10.1007/sll263-015-0816-y
31
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
155.	Krizhevsky A., Hinton G., 2009: Learning multiple layers of features from tiny images.
Available online at: http://www.cs.utoronto.ca/~kriz/learning-features-2009-TR.pdf (Accessed
6July2023).
156.	Neverova N., Wolf C., Lacey G., Fridman L., Chandra D., Barbello B., Taylor G., 2016:
Learning human identity from motion patterns. IEEE Access, 4, 1810-1820.
https://doi.org/10.1109/access.2016.2557846
157.	Răni C.J., Devarakonda N., 2022: An effectual classical dance pose estimation and
classification system employing convolution neural network-long shortterm memory (CNN-
LSTM) network for video sequences. Microprocessors and Microsystems, 95, 104651.
https://doi.Org/10.1016/j.micpro.2022.104651
158.	Vaith A., Taetz B., Bleser G., 2020: "Uncertainty Based Active Learning with Deep Neural
Networks for Inerțial Gait Analysis," in 2020 IEEE 23rd International Conference on Information
Fusion (FUSION), 1-8. https://doi.org/10.23919/fusion45008.2020.9190449
159.	Hu Q., Tang X. and Tang W., 2020: A real-time patient-specific sleeping posture
recognition system using pressure sensitive conductive sheet and transfer learning. IEEE
Sensors Journal, 21(5), 6869-6879. https://doi.org/10.1109/JSEN.2020.3043416
160.	Long C., Jo E., Nam Y., 2022: Development of a yoga posture coaching system using an
interactive display based on transfer learning. The Journal of Supercomputing, 1-16.
https://doi.org/10.1007/sll227-021-04076-w
161.	Krizhevsky A., Sutskever L, Hinton G.E., 2012: Imagenet classification with deep
convolutional neural networks. Advances in neural Information processing Systems. Available
online at: https://proceedings.neurips.cc/paper/2012/file/c399862d3b9d6b76c8436e924a68c45b-
Paper.pdf (Accessed 28 June 2023).
162.	Zhang Y., Huang Y., Yu S., Wang L., 2019: Cross-view gait recognition by discriminative
feature learning. IEEE Transactions on Image Processing, 29,	1001-1015.
https://doi.org/10.1109/tip.2019.2926208
163.	Newell A., Yang K., Deng J., 2016: Stacked hourglass networks for human pose
estimation. In Computer Vision-ECCV 2016: 14th European Conference, Amsterdam, The
Netherlands, October 11-14, 2016, Proceedings, Part VIII 14, 483-499. Springer International
Publishing. https://doi.org/10.1007/978-3-319-46484-8_29
164.	Chen Y., Wang Z., Peng Y., Zhang Z., Yu G., Sun J., 2018: Cascaded pyramid network for
multi-person pose estimation. In Proceedings of the IEEE conference on computer vision and
pattern recognition, 7103-7112. https://doi.org/10.1109/CVPR.2018.00742
165.	Wei S.E., Ramakrishna V., Kanade T., Sheikh, Y., 2016: Convolutional pose machines. In
Proceedings of the IEEE conference on Computer Vision and Pattern Recognition, 4724-4732.
https://doi.org/10.48550/arXiv.1602.00134
166.	Sun K., Xiao B., Liu D., Wang J., 2019: Deep high-resolution representation learning for
human pose estimation. In Proceedings of the IEEE/CVF conference on computer vision and
pattern recognition, 5693-5703. https://doi.org/10.1109/CVPR.2019.00584
32
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
167.	Redmon J., Divvala S., Girshick R., Farhadi A., 2016; You only look once: Unified, real-
time object detection. In Proceedings of the IEEE conference on computer vision and pattern
recognition, 779-788. https://doi.org/10.48550/arXiv.1506.02640
168.	Lee M.F.R., Chen Y.C., Tsai C.Y., 2022: Deep learning-based human body posture
recognition and tracking for unmanned aerial vehicles. Processes, 10(11), 2295.
https://doi.org/10.3390/prl0112295
169.	Cao Z., Simon T., Wei S.E., Sheikh Y., 2017: Realtime multi-person 2d pose estimation
using part affinity fields. In Proceedings of the IEEE conference on computer vision and pattern
recognition, 7291-7299. https://doi.org/10.48550/arXiv.1611.08050
170.	Li B., Han C., Bai B., 2019: Hybrid approach for human posture recognition using
anthropometry and BP neural network based on Kinect V2. EURASIP Journal on Image and
Video Processing, 2019, 1-15. https://doi.org/10.1186/sl364 0-018-0393-4
171.	Chen K., Wang Q., 2015, December: Human posture recognition based on skeleton data.
In 2015 IEEE International Conference on Progress in Informatics and Computing (PIC), 618-
622. ieee. https://doi.org/10.1109/PIC.2015.7489922
172.	Chollet F., 2017: Xception: Deep learning with depthwise separable convolutions. In
Proceedings of the IEEE conference on computer vision and pattern recognition, 1251-1258.
https://doi.org/10.48550/arXiv.1610.02357
173.	Li Y., Huang H., Xie Q., Yao L., Chen Q., 2018: Research on a surface defect detection
algorithm based on MobileNet-SSD. Applied Sciences, 8(9),	1678.
https://doi.org/10.3390/app8091678
174.	Zhang X., Zhou X., Lin M., Sun J., 2018: Shufflenet: An extremely efficient convolutional
neural network for mobile devices. In Proceedings of the IEEE conference on computer vision
and pattern recognition, 6848-6856. https://doi.org/10.48550/arXiv.1707.01083
175.	Ma N., Zhang X., Zheng H.T., Sun, J., 2018: Shufflenet v2: Practicai guidelines for efficient
cnn architecture design. In Proceedings of the European conference on computer vision
(ECCV), 116-131. https://doi.org/10.48550/arXiv.1807.11164
176.	Zhang J., Zhu H., Wang P., Ling X., 2021: ATT squeeze U-Net: A lightweight network for
forest fire detection and recognition. IEEE Access, 9,	10858-10870.
https://doi.org/10.1109/ACCESS.2021.3050628
177.	Dindorf C., Ludwig O., Simon S., Becker S., Frdhlich M., 2023: Machine learning and
explainable artificial intelligence using counterfactual explanations for evaluating posture
parameters. Bioengineering, 10(5), 511. https://doi.org/10.3390/bioengineeringl0050511
178.	Alpaydin E., 2020: Introduction to machine learning. MIT press. Available online at:
http://nuir.nkumbauniversity.ac.ug/xmlui/bitstream/handle/20.500.12383/1421/Introduction%
20to%20Machine%20Learning,0/o20Second<7o20Edition%20(Adaptive%20Computation%20and
%20Machine%20Learning)%20(%20PDFDrive%20).pdf?sequence=l (Accessed 19 June 2023).
179.	LeCun Y., Bengio Y., Hinton G., 2015: Deep learning. Nature, 521(7553), 436-444.
https://doi.org/10.1038/naturel4539
33
Revista Pădurilor 138(3) (2023) 001-034
Forkuo: A systematic survey on convențional and new postural assessment methods
180.	Anaconda Inc., 2023: The definitive guide to AI Platforms for open-source data Science
and ML. Available online at: https://www.anaconda.com/definitive-guide-to-ai-platforms
(Accessed 21 June 2023).
181.	Hinton G., Deng L., Yu D., Dahl G.E., Mohamed A.R., Jaitly N., Senior A., Vanhoucke V.,
Nguyen P., Sainath T.N., Kingsbury B., 2012: Deep neural networks for acoustic modeling in
speech recognition: The shared views of four research groups. IEEE Signal processing
magazine, 29(6), 82-97. https://doi.org/10.1109/MSP.2012.2205597
182.	Currenti J., Styduhar C., Yang C., 2023: The definitive guide to AI platforms for open-
source data Science and ML. Available online at: https://www.anaconda.com/definitive-guide-
to-ai-platforms (Accessed 20 June 2023).
34
Revista Pădurilor 138(3) (2023) 001-044
Revista Pădurilor
disponibil online la: www.revistapadurilor.com
DOMNUL PROFESOR DR. ING. ARCADIE CIUBOTARU
ÎMPLINEȘTE 75 DE ANI
9
Elena Camelia Musata*
9
aDepartamentul de Exploatări Forestiere, Amenajarea Pădurilor și Măsurători Terestre, Facultatea de
Silvicultură și exploatări forestiere, Universitatea Transilvania din Brașov, Șirul Beethoven 1, 500123,
Brașov, România, e-mail: elena.musat@unitbv.ro (E.C.M.).
REPERE
•	Aniversarea domnului
prof. dr. ing. Arcadie
Ciubotaru la venerabila
vârstă de 75 ani.
REZUMAT GRAFIC
INFORMAȚII ARTICOL
Istoricul articolului:
Manuscris primit la: 4 septembrie 2023
Primit în forma revizuită: -
Acceptat: 4 septembrie 2023
Număr de pagini: 4 pagini.
Tipul articolului:
Comunicare
Editor: Stelian Alexandru Borz
Cuvinte cheie:
Arcadie Ciubotarii
Aniversare
75 de ani
REZUMAT
Născut în 12 august 1948 în Buhuși, județul Bacău, domnul prcfesor a fost mereu
apropiat de natură și de tradiții, care l-au definit și l-au ghidat în dezvoltarea
profesională și personală. Astfel, între 1955 - 1962 a urmat cursurile la Școala
Generală din satul Slobozia, județul Neamț, după care, în 1966, tânărul Arcadie
Ciubotaru a absolvit Liceul Teoretic Roznov, din același județ. Apropierea de
natură, ca urmare a locului unde și-a petrecut primii ani din viață, l-a influențat
pe domnul prcfesor în alegerea profesiei pe care să o urmeze și căreia să i se dedice
întreaga viață. Așa că, în 1968 a dat admitere la Facultatea de Silvicultură din
Brașov, iar în 1973 i s-a îndeplinit visul de a deveni inginer silvic. Pe parcursul
celor 35 de ani în care a activat ca și cadru didactic, domnul prcfesor a pregătit
generații și generații de ingineri silvici, prezentându-le cu măiestrie și dăruire
tainele exploatării și prelucrării primare a lemnului. Privind per ansamblu
întreaga carieră a domniei sale, se poate observa cu ușurință că valorile pe care a
pus accent întotdeauna domnul prcfesor au fost profesionalism, seriozitate,
implicare, organizare și, nu în ultimul rând, omenie. Vă mulțumim domnule
prcfesor pentru tot ceea ce ne-ați oferit nouă, studenților, de la cunoștințe și
îndrumări, până la sfaturi părintești, și vă dorim mulți ani înainte, cu sănătate,
liniște și nenumărate bucurii alături de persoane dragi dumneavoastră!
* Autor corespondent. Tel.: +40-761-646-361.
Adresa de e-mail: elena.musat@unitbv.ro
Revista Pădurilor 138(3) (2023) 035-038
Mușat: Domnul profesor dr. ing. Arcadie Ciubotaru împlinește 75 de ani
DOMNUL PROFESOR DR. ING. ARCADIE CIUBOTARU
ÎMPLINEȘTE 75 DE ANI
Un om deosebit ne-a spus odată nouă, studenților, că „este mare lucru să mergi la muncă cu
drag" și că „atâta timp cât tu nu ești perfect, nu-i cere celuilalt să fie". Acest om este domnul profesor
Arcadie Ciubotaru, care întotdeauna s-a ghidat după principii solide, bazate atât pe experiențele
proprii, cât și pe cunoștințele transmise de-a lungul timpului, din moși-strămoși.
Născut în 12 august 1948 în Buhuși, județul Bacău, domnul profesor a fost mereu apropiat de
natură și de tradiții, care l-au definit și l-au ghidat în dezvoltarea profesională și personală. Astfel,
între 1955-1962 a urmat cursurile la Școala Generală din satul Slobozia, județul Neamț, după care, în
1966, tânărul Arcadie Ciubotaru a absolvit Liceul Teoretic Roznov, din același județ (1962-1966).
Apropierea de natură, ca urmare a locului unde și-a petrecut primii ani din viață, l-a influențat
pe domnul profesor în alegerea profesiei pe care să o urmeze și căreia să i se dedice întreaga viață.
Așa că, în 1968 a dat admitere la Facultatea de Silvicultură din Brașov, iar în 1973 i s-a îndeplinit
visul de a deveni inginer silvic.
Ca urmare a politicilor vremii, după absolvirea facultății din Brașov, tânărul inginer a fost
repartizat și mai departe de casă, la G.S.C.F.I. Râmnicu Vâlcea unde, timp de trei ani (1973-1976), și-
a îmbunătățit cunoștințele dobândite în facultate, având posibilitatea de a le aplica practic, în
producție, alături de alți silvici „cu state vechi" de la care a continuat să învețe, ghidându-se pe ideea
că „de la oricine poți învăța câte ceva dacă ai mintea deschisă".
După perioada de așa-zisă „ucenicie" în ceea ce privește profesia de inginer silvic, domnul
profesor Ciubotaru și-a continuat activitatea la I.F.E.T. Orșova, dar de această dată din postura de
șef de lot și specialist silvic, nu de învățăcel. în cei doi ani (1976-1978) a trebuit să facă față multor
provocări pe care, prin cunoștințele însușite, seriozitate și organizare, a reușit să le ducă la bună
îndeplinire. Aici, postul ocupat a impus nu numai activități care țineau de exploatarea lemnului, ci
și activități de proiectare a drumurilor forestiere, de supraveghere a construcției acestora și multe
alte sarcini pe care domnul profesor le-a îndeplinit cu brio, învățând din fiecare câte ceva și
îmbunătățindu-și permanent cunoștințele și aptitudinile tehnice și organizatorice.
Cu toate frumusețile din zonă și tradițiile păstrate încă de localnici și împărtășite cu drag noilor
veniți, inginerul Arcadie Ciubotaru a resimțit lipsa Brașovului și, mai ales, a familiei întemeiate aici.
Astfel, în 1978 a revenit în orașul de la poalele Tâmpei, unde și-a dedicat timpul, cunoștințele și
experiența acumulată studenților de la Facultatea de Silvicultură.
Domnul profesor a parcurs, din aproape în aproape, toate treptele ierarhiei universitare, de la
asistent universitar (1978-1991), la șef de lucrări (1991-1995), conferențiar universitar (1995-1998) și,
în final, profesor universitar (1998-2013), predând ore la disciplinele Exploatarea pădurilor (curs și
laborator - licență, programul de studii Silvicultură, anul III), Tehnologia prelucrării lemnului (curs
și laborator - licență, programul de studii Exploatări forestiere, anul IV), Proiectarea și organizarea
activității de exploatare (curs și proiect - licență, programul de studii Cinegetică ID, anul III),
36
Revista Pădurilor 138(3) (2023) 035-038
Mușat: Domnul profesor dr. ing. Arcadie Ciubotaru împlinește 75 de ani
Ecotehnologii de exploatare a pădurilor (curs și laborator - masterat, programul de studii
Management și sisteme tehnice în exploatări forestiere, anul I).
Pe parcursul celor 35 de ani în care a activat ca și cadru didactic, domnul profesor a pregătit
generații și generații de ingineri silvici, prezentându-le cu măiestrie și dăruire tainele exploatării și
a prelucrării primare a lemnului. Fiind un adevărat profesionist, care iubește natura și activitățile
practice, domnul profesor a organizat numeroase deplasări în teren astfel încât studenții să poată
observa cum se recoltează si se colectează efectiv lemnul si chiar să mânuiască ferăstrăul mecanic.
Tactul didactic și expunerile interesante făceau ca la cursurile domnului profesor sălile să fie
aproape pline, chiar dacă nu făcea niciodată prezența. în plus, sfaturile, învățăturile și glumele
presărate din loc în loc, întotdeauna legate de subiectul abordat, ne făceau pe noi, studenții, să-1
îndrăgim și mai mult și să fim mai motivați în a deprinde noi și noi cunoștințe. Profesionalismul
demonstrat la fiecare oră de curs și lucrări practice, precum și personalitatea domnului profesor au
făcut ca mulți dintre studenți să aleagă să realizeze proiectul de diplomă și/sau lucrarea de disertație
sub îndrumarea dânsului.
Aprecierea generațiilor de studenți îndrumați în lucrările de finalizare a studiilor sau care l-au
avut pe domnul profesor doar cadru didactic pe parcursul studiilor, s-a extins și în exteriorul
instituției, motiv pentru care domnul profesor Ciubotaru este întâmpinat cu zâmbetul pe buze și cu
mare drag oriunde merge în țară. Aceste aprecieri se traduc și prin numeroasele solicitări din partea
absolvenților de a realiza fie teze de doctorat sub îndrumarea dumnealui, fie lucrări de grad didactic
I în învățământul preuniversitar.
în acest sens, domnul profesor a dobândit dreptul de a conduce doctorat în anul 1991. Această
calitate presupune multe provocări, atât prin prisma seriozității celor înscriși la doctorat, cât și a
subiectelor abordate. Astfel, domnul profesor a dat dovadă de spirit organizatoric, gândire logică și
o dorință continuă de perfecționare, care l-au condus la îndrumarea cu succes a mai multor teze de
doctorat care au abordat fie teme specifice exploatării lemnului (consumuri și pierderi tehnologice
la exploatarea arborilor în diverse condiții de relief și arboret, calitatea arborilor pe picior în arborete
37
Revista Pădurilor 138(3) (2023) 035-038
Mușat: Domnul profesor dr. ing. Arcadie Ciubotaru împlinește 75 de ani
pure de fag și arborete tinere de molid, caracteristicile defectelor vizibile la arborii de fag), fie teme
apropiate acestui domeniu (calitatea arborilor din zonele publice) și chiar teme care au presupus
provocări mai mari, precum utilizarea tehnologiei LiDAR în realizarea modelului digital altimetric
al terenurilor acoperite cu vegetație forestieră.
Remarcându-se în cadrul colegilor ca o persoană organizată, cu gândire logică, perseverentă
și dornică de implicare, domnul profesor a fost desemnat și în diverse poziții de conducere de la
nivelul universității, ocupând funcțiile de director educativ, director de colegiu universitar și
cancelar general al Universității Transilvania din Brașov. Dar, cu toate aceste noi responsabilități,
domnul profesor nu și-a neglijat niciodată activitatea de bază, nici pe cea didactică și nici pe cea
științifică. Din această perspectivă a publicat numeroase cărți și îndrumare de laborator pentru
studenți, pe specificul disciplinelor predate. în plus, preocupările din cercetare s-au concretizat în
31 de proiecte de cercetare științifică, cu teme solicitate de diverse instituții din țară. De asemenea,
trebuie menționate și cele peste 130 de articole publicate în reviste de specialitate, unele dintre ele
fiind prezentate și la conferințe internaționale, iar altele publicate în reviste cotate la cel mai înalt
nivel.
în 2013, la vârsta de 65 de ani, domnul profesor Arcadie Ciubotaru s-a pensionat, păstrându-
și calitatea de cadru didactic asociat și predând ore la disciplinele Exploatarea pădurilor și
Ecotehnologii de exploatare a pădurilor.
La 70 de ani a renunțat și la această calitate, păstrându-și doar statutul de conducător de
doctorat. în această calitate, domnul profesor continuă să studieze, să se perfecționeze și le oferă încă
îndrumare, idei și sfaturi atât doctoranzilor, cât și celor pe care i-a îndrumat, dar care și-au găsit un
loc în domeniul învățământului sau a cercetării științifice, domenii pe care domnul profesor le-a
onorat toată viața și continuă să o facă.
Privind per ansamblu întreaga carieră a domniei sale, se poate observa cu ușurință că valorile
pe care a pus accent întotdeauna domnul profesor au fost profesionalism, seriozitate, implicare,
organizare și, nu în ultimul rând, omenie. Aceste calități l-au ajutat pe domnul profesor Arcadie
Ciubotaru să devină un pedagog deosebit, un prieten și un coleg de nădejde, dar, mai ales, un model
de viață și profesionalism pentru mulți dintre noi.
Domnule profesor, pentru mine și mulți alții sunteți dovada vie că poți merge la muncă cu
drag și că numai atunci când faci ceva cu plăcere vei obține și rezultatele așteptate. Vă mulțumim
domnule profesor pentru tot ceea ce ne-ați oferit nouă, studenților, de la cunoștințe și îndrumări,
până la sfaturi părintești, și vă dorim mulți ani înainte, cu sănătate, liniște și nenumărate bucurii
alături de persoane dragi dumneavoastră!
La mulți ani!!!
Ș.L. dr. ing. Elena Camelia MUȘAT
Facultatea de Silvicultură și exploatări forestiere
Universitatea Transilvania din Brașov
38
Revista Pădurilor 138(3) (2023) 001-044
Revista Pădurilor
disponibil online la: www.revistapadurilor.com
PROF. DR. BIOL. MARIUS DANCIU (1941-2023), UN BOTANIST
ÎNTRE SILVICULTORI
Dan-Marian Gureana'*, Pantelimon Ularua, Victor Adrian Indreicaa
aDepartamentul de Silvicultură, Facultatea de Silvicultură și exploatări forestiere, Universitatea
Transilvania din Brașov, Șirul Beethoven 1,500123, Brașov, România.
, z ,	r	r	, r
REZUMAT GRAFIC
REPERE
•	Cadru didactic titular la
Facultatea de Silvicultură și
exploatări forestiere din
Brașov, disciplina Botanică
și Fiziologia plantelor, în
perioada 1964-2006.
•	108 lucrări științifice (16
cărți, 92 articole), 27 de
contracte de cercetare.
•	cercetări floristice - două
specii noi pentru flora
României identificate,
cercetări fitocenologice.
INFORMAȚII ARTICOL
Istoricul articolului:
Manuscris primit la: 4 septembrie 2023
Primit în forma revizuită: -
Acceptat: 4 septembrie 2023
Număr de pagini: 6 pagini.
Tipul articolului:
Comunicare
Editor: Stelian Alexandru Borz
Cuvinte cheie:
Profesor Marius Danciu
Activitatea didactică și
științifică
Facultatea de Silvicultură
și exploatări forestiere
REZUMAT
Absolvent al secției de biologie-botanică a Facultății de Biologie-Geogrcfie din
cadrul Universității „Babeș-Bolyai" din Cluj-Napoca, dl. Prcfesor Marius
Danciu a beneficiat de repartiție ministerială în învățământul superior, fiind
angajat ca șef de laborator în octombrie 1964 la Disciplina Botanică de la
Facultatea de Silvicultură și exploatări forestiere din Brașov. A devenit apoi
botanist, botanist principal, asistent (1972), șef de lucrări (1990), conferențiar
(1993) și prcfesor universitar din 1998. A susținut doctoratul în geobotanică la
Universitatea București, sub conducerea Prcf dr. doc. biol. luliu Morariu, cu
teza „Studii geobotanice în sudul Munților Baraolt". A publicat 108 lucrări
științifice (16 cărți, 4 ca prim autor și 92 de articole, 11 ca unic autor), a
contribuit la elaborarea a 27 de contracte de cercetare ca responsabil sau
colaborator, circumscrise domeniilor botanicii - cercetări floristice, fitocenologice,
fenologice, eccfiziologice, privind corologia, ecologia și fitocenologia speciilor de
cormcfite, resursele de plante medicinale din fondul forestier sau cele trcfice
pentru vânatul ierbivor din diferite ecosisteme, conservarea speciilor sau
comunităților rare de plante.
* Autor corespondent. Tel.: +40-740-080-430.
Adresa de e-mail: dangurean@unitbv.ro
Revista Pădurilor 138(3) (2023) 039-044
Gurean et al.: Prof. dr. biol. Marius Danciu (1941 - 2023)...
PROF. DR. BIOL. MARIUS DANCIU (1941-2023), UN
BOTANIST ÎNTRE SILVICULTORI
Domnul Prof. dr. biol. Marius Danciu s-a născut pe 16 august 1941, în satul Camna (comuna
Tauț, în prezent comuna Șilindia, județul Arad), în casa părinților săi - învățătorul Traian Danciu și
mama - Mărioara Danciu, casnică, dedicată treburilor casei și creșterii copiilor - Lavinia și Marius,
care au moștenit de la tatăl lor vocația profesiei de dascăl și au dus mai departe această frumoasă
tradiție de familie, ca învățătoare în Arad, și respectiv profesor universitar la Brașov.
Cursurile școlii elementare le-a urmat în perioada 1948-1955, clasele I-II în satul natal, clasele
III-IV în satul Chier (comuna Târnova, județul Arad) iar la numai 11 ani a plecat de acasă pentru a-
și continua studiile cu clasele V-VII (1952-1955) în orașul Ineu (județul Arad), unde a urmat și
cursurile liceale la „Școala Medie Mixtă", secția reală, între 1955 și 1959.
în perioada 1959-1964, domnul Profesor Danciu a frecventat cursurile secției de biologie-
botanică a Facultății de Biologie-Geografie din cadrul Universității „Babeș-Bolyai" din Cluj-Napoca,
obținând titlul de diplomat (licențiat) în biologie, specialitatea botanică. Pe baza rezultatelor obținute
în anii de studii universitare, a beneficiat de repartiție ministerială într-o instituție de învățământ
superior, fiind încadrat pe 28.10.1964 ca șef de laborator (funcție ce corespundea atunci celei de
preparator universitar), la Disciplina Botanică din structura Catedrei de Botanică și Zoologie
forestieră a Facultății de Silvicultură și exploatări forestiere, din cadrul Institutului Politehnic
Brașov. Șeful catedrei - Prof. dr. doc. biol. luliu Morariu, solicitase repartizarea unui biolog-botanist
pe postul menționat, devenit vacant ca urmare a transferului domnului dr. biolog Pantelimon Ularu,
ca asistent universitar, la nou înființata Facultate de Științe Naturale din cadrul Institutului
Pedagogic de 3 ani al Universității din Brașov.
în perioada 1965-1969 domnul Profesor Danciu a deținut în cadrul aceleiași discipline funcția
de botanist, apoi, din 1969, odată cu modificarea titulaturii disciplinei, botanist principal la Botanică
și fiziologia plantelor, asistent suplinitor în anul universitar 1971-1972 concomitent cu preluarea
poziției de titular de către Prof. dr. ing. biol. Darie Parascan iar din 01.10.1972, este titularizat prin
concurs, când acumulase deja 8 ani vechime, pe post de asistent universitar - poziția 6 în Statul de
funcții al noii Catedre de Silvicultură, constituită prin includerea, alături de cea veche, a catedrelor
desființate de Botanică și Zoologie forestieră și a celei de împăduriri și Ameliorații silvice. Domnul
Profesor Danciu s-a integrat cu succes în noul colectiv de catedră, dominat evident de silvicultori, în
timp ce biologii alături de care își începuse activitatea s-au retras, pe rând, prin pensionare (Prof. dr.
doc. biol. luliu Morariu, pensionat în 1970), transfer (dr. biol. Pantelimon Ularu), sau ca urmare a
plecării din țară (dr. biol. Heinz Heltmann).
Colegii mai vechi sau noi, dar și zeci de promoții de studenți ce l-au cunoscut, au apreciat
unanim și permanent calitățile remarcabile ce au definit personalitatea Profesorului Marius Danciu
- ținuta morală ireproșabilă, onestitatea, modestia, delicatețea, dedicarea, meticulozitatea, rigoarea
științifică, conștiinciozitatea și, nu în ultimul rând, tenacitatea.
40
Revista Pădurilor 138(3) (2023) 039-044
Gurean et al.: Prof. dr. biol. Marius Danciu (1941 - 2023)...
în 1966 domnul profesor și-a unit destinul cu cel al d-rei Knobloch Catalcna, absolventă a
aceleiași universități clujene - Facultatea de Filologie, secția limbă și literatură maghiară, profesoară
la Covasna, apoi corector, translator, redactor și secretar de redacție la ziarul „Brassoi Lapok" din
Brașov, căsătoria lor fiind binecuvântată cu un copil - Daniel, născut în 1975.
Toate calitățile menționate ale domnului profesor, credința fermă a domniei sale că munca
asiduă, conștiinciozitatea de excepție, meritele și rezultatele incontestabile vor trebui să fie la un
moment dat recompensate, dar și susținerea permanentă a familiei l-au ajutat să depășească
parcursul nefiresc de lung și întru totul nemeritat al treptelor ierarhiei didactice, o primă și
importantă neîmplinirc a carierei domniei sale.
După 12 ani de activitate ca asistent și alți 8 anterior ca șef de laborator, botanist și botanist
principal, timp în care a susținut doar ore de lucrări practice, în 1984 domnul Danciu preia cursul
de Botanică și fiziologia plantelor de la anul I ingineri seral și doi ani mai târziu părea să apară
oportunitatea unei promovări, conform adeverinței nr. 8851, eliberată de Universitatea din Brașov
la data de 25.11.1986, spre a-i servi la completarea dosarului în vederea susținerii concursului pentru
ocuparea postului de Șef de lucrări, poziția 18 în Statul de funcții al Catedrei de Silvicultură. Din
păcate, din motive pe care nu le cunoaștem, acel concurs nu a mai avut loc sau a fost amânat, așa că
dl. Danciu a devenit Șef de lucrări abia după Revoluția din decembrie 1989, mai exact de la 1.10.1990,
la 26 de ani vechime, când în Fișa de prezentare din 27.02.1990 a candidatului - Asist, dr. biol. Marius
Danciu, la concursul pentru postul de Șef de lucrări, este consemnat că a publicat singur sau în
colaborare 61 de lucrări științifice, este doctor în biologie din 1974, a participat în calitate de
responsabil sau colaborator la elaborarea a 19 contracte de cercetare.
Ca semn al unei relative normalizări a promovărilor, și al fundamentării acestora
preponderent pe criterii obiective de performanță științifică și didactică, dl. Danciu este titularizat,
prin concurs, pe post de conferențiar universitar, începând cu 1.10.1993, din 1994 devine titularul
cursului de Botanică, succedându-i Prof. dr. ing. biol. Darie Parascan și, la 34 de ani vechime în
învățământul superior, devine profesor universitar, de la 01.10.1998, deși întrunea condițiile
necesare în acest sens de multă vreme.
A funcționat pe acest post până în 2006, dar din 2005 s-a confruntat cu a doua neîmplinire din
carieră, și anume reducerea, prin Reforma Bologna și restrângerea învățământului superior tehnic
de la 5 la 4 ani, de la un an întreg la un singur semestru a timpului acordat cursului și lucrărilor
practice de Botanică în Planul de învățământ al programelor de studii Silvicultură, Exploatări
forestiere și Cinegetică, situație care le-a creat domnilor profesori Danciu și Parascan o mare
dezamăgire. Nu a trecut mult și, din păcate, în 2006 Profesorul Danciu trăiește nemeritat o a treia
neîmplinire - imposibilitatea de a obține conducerea de doctorat în domeniul fundamental Științe
agricole și silvice, domeniul Ecologie forestieră sau Botanică forestieră, datorită faptului că era
doctor în biologie, specializarea geo-botanică. Nu a contat faptul că, deși biolog, cu doctoratul în
domeniul de formare, și-a desfășurat activitatea timp de 42 de ani preponderent între silvicultori,
nici că a făcut eforturi majore - în continuarea celor ale predecesorilor săi, Prof. dr. ing. biol. Darie
Parascan, și Prof. dr. doc. biol. luliu Morariu, pentru a adapta conținutul cursului și lucrărilor
practice de Botanică specificului forestier, în sensul exemplificării termenilor consacrați de
morfologie, ori de câte ori a fost posibil, la specii lemnoase de interes forestier, al realizării de
preparate microscopice cu structurile diferitelor organe vegetative sau reproducătoare tot la specii
lemnoase sau ierboase de pădure, dar și în sensul focusării în cadrul sistematicii asupra familiilor și
41
Revista Pădurilor 138(3) (2023) 039-044
Gurean et al.: Prof. dr. biol. Marius Danciu (1941 - 2023)...
ordinelor care includ plante ierboase indicatoare de pădure, deci specii cu evidente afinități pentru
mediul forestier. Nu au contat nici recomandările, oferite cu generozitate, de către Prof. consult, dr.
ing. biol. dr. Honoris Causa Darie Parascan, membru titular A.S.A.S. și șef al Catedrei de Silvicultură
în perioada 1985-1990, cu care dl. Prof. Danciu a colaborat practic pe întreg parcursul activității la
catedră, Prof. dr. ing. Dumitru Târziu, membru titular A.S.A.S, șef al Catedrei de Silvicultură în
perioada 1990-2004 dar și Prof. consult, dr. biol. Vasile Ciocîrlan de la Universitatea de Științe
Agronomice și Medicină Veterinară din București, Catedra de Botanică și Fiziologie vegetală.
Informat că nu poate primi conducerea de doctorat decât în biologie - geo-botanică, ceea ce
evident nu îl avantaja în contextul activității domniei sale la Facultatea de Silvicultură și exploatări
forestiere, ai cărei absolvenți nu aveau posibilitatea de a se înscrie la un astfel de doctorat, domnul
Profesor Danciu a ales să nu își mai depună dosarul și, în consecință, conform reglementărilor
momentului, nu a putut să își prelungească activitatea în calitate de titular, fiind nevoit să se
pensioneze la împlinirea vârstei de 65 de ani, în 2006. A mai predat cursul de Botanică până în anul
2007 și apoi, timp de câțiva ani a mai ajutat disciplina prin susținerea unor ore de lucrări practice în
regim de plata cu ora sau chiar pe parcursul practicii, după care, cu toate rugămințile noastre, s-a
retras din activitatea didactică, motivând, pe bună dreptate, dezinteresul tot mai accentuat al multor
studenți față de școală, lipsa de motivare a acestora pentru o pregătire profesională temeinică.
Activitatea științifică a d-lui Profesor Danciu a fost orientată mai întâi, după cum însuși
dumnealui afirma într-un memoriu de activitate, „cu precădere asupra cercetărilor floristice,
materializată îndeosebi în dezvoltarea și continua perfectare, prin verificare critică și punerea de
acord cu lucrările taxonomice moderne, a colecției științifice de plante presate a facultății".
în perioada 1964-1990, „stocului comparativ de materiale științifice ale herbarului i s-au
adăugat 25.600 de coli noi, reprezentând taxoni (specii, subspecii, varietăți) din diferite unități
sistematice", în această cifră fiind incluse specimene recoltate și de ceilalți membri ai colectivului
disciplinei sau chiar de întregul colectiv pe parcursul unor deplasări (de exemplu pe Măgura
Codlei), dar și materialele herbaristice obținute prin schimburi cu alte colecții din țară (Institutul de
Biologie din București, Universitățile din Iași, Cluj-Napoca, Craiova, Institutul Pedagogic din Bacău
etc.) sau din străinătate (Muzeul din Lund, Suedia). Fără îndoială însă, profesorul Marius Danciu a
contribuit în cea mai mare măsură la îmbogățirea Herbarului Facultății de Silvicultură și exploatări
forestiere din Brașov și a investit extrem de mult timp în verificarea atentă a taxonilor herborizați,
mai recent sau cu mai multă vreme în urmă, ceea ce l-a determinat să afirme, riguros fundamentat,
că „este unul din cele mai bogate și mai bine sistematizate din țară". Ținând cont de acestea,
apreciem că ar fi pe deplin meritat ca acest herbar să-i poarte numele, și vom face toate demersurile
necesare în acest sens.
Activitatea publicistică a domnului profesor a debutat în 1967 cu articolul Vegetația
stâncăriilor de pe Măgura Codlei, autori Morariu L, Ularu P., Danciu, M., Lungescu E., publicat în
Buletinul Institutului Politehnic Brașov, 9, pg. 15-24. Din cele 108 lucrări științifice, 16 (4 ca prim
autor) sunt cărți, capitole în cărți sau îndrumare de lucrări practice, publicate în edituri centrale ca
Editura „Ceres" - Morfologia și fiziologia plantelor lemnoase, cu elemente de taxonomie vegetală,
1983, 363 p., autori Parascan D., Danciu M., care a primit premiul „Traian Săvulescu" al Academiei
Republicii Socialiste România sau Edit. Silvică - Plante vasculare periclitate, vulnerabile și rare din
pădurile României, 2007, 258 p., autori Danciu M., Gurean D., Indreica A., indexată în baza
internațională de date UNPD (United Nations Programs for Development) Romania, edituri locale
42
Revista Pădurilor 138(3) (2023) 039-044
Gurean et al.: Prof. dr. biol. Marius Danciu (1941 - 2023)...
ca Editura „Pentru Viață", Brașov - Botanică forestieră, 2002, 324 p., autori Danciu M., Parascan D.,
Gurean D. (colaborator), Editura și Reprografia Universității „Transilvania" din Brașov. Din cele 92
de articole științifice publicate în diverse reviste sau periodice de prestigiu din țară sau din
străinătate, în volumele unor conferințe sau sesiuni de comunicări științifice, 11 îl au pe dl. Prof.
Danciu ca unic autor, 16 au fost elaborate în colectivul de biologi alături de care și-a început
activitatea (dr. P. Ularu - prim autor la 3 dintre acestea, dr. E. Lungescu), sub coordonarea Prof. dr.
doc. luliu Morariu, 23 în colective mai largi din care a făcut parte și Prof. dr. ing. biol. Darie Parascan,
21 în colaborare numai cu dânsul, fiind alternată coordonarea.
Studiile doctorale le-a început în 1968, în specialitatea geobotanică, sub îndrumarea Prof. dr.
doc. Ion Popescu-Zeletin, membru corespondent al Academiei Române, și le-a continuat după
decesul acestuia în 1974, sub conducerea Prof. dr. doc. luliu Morariu. A susținut la Universitatea
București, pe data de 23.12.1974, teza de doctorat „Studii geobotanice în sudul Munților Baraolt",
obținând în același an titlul științific de doctor în biologie.
Lucrările de cercetare științifică ale domnului Prof. dr. biol. Marius Danciu se circumscriu
diferitelor domenii ale botanicii - cercetări floristice, fitocenologice, referitoare la corologia, ecologia
și fitocenologia speciilor de cormofite, conservarea speciilor sau comunităților rare de plante,
cunoașterea resurselor trofice pentru vânatul ierbivor din diferite ecosisteme, cercetări fenologice,
ecofiziologice sau privind cunoașterea resurselor de plante medicinale din fondul forestier, precum
și a diferitelor erbicide și a eficienței acestora în culturile din pepiniere sau plantații. Rezultatele
cercetărilor sale s-au concretizat în identificarea a două specii noi pentru Flora României (de ex.
Geranium cinereum, semnalată în 1983 de dl. profesor de pe stâncării foarte greu accesibile din Cheile
Postăvarului - Figura 1), numeroase stațiuni noi de răspândire a unor taxoni rari sau sporadici,
precum și prin contribuții remarcabile la vegetația anumitor teritorii, masive muntoase sau chiar, în
ansamblu, la vegetația României.
Figura 1. Geranium cinereum pe stâncării, la Porțile de Piatră ale Postăvarului. Foto: Gurean D.
43
Revista Pădurilor 138(3) (2023) 039-044
Gurean et al.: Prof. dr. biol. Marius Danciu (1941 - 2023)...
O parte din aceste rezultate au rămas din păcate nepublicate, dar ne vom strădui ca ele să vadă
lumina tiparului, cum este și cazul unui articol aflat acum în curs de publicare - Contributions to the
flora and vegetation of Piatra Craiului National Park (I), autori Danciu M., Gurean D., Indreica A.,
comunicare susținută la Sesiunea științifică bienală „Forest and sustainable development",
organizată de Facultatea de Silvicultură și exploatări forestiere din Brașov în octombrie 2022.
Domnul Prof. Danciu a participat în calitate de colaborator sau de responsabil la elaborarea a
27 de contracte de cercetare cu diferiți beneficiari - Regia Națională a Pădurilor ROMSILVA, I.C.A.S.
București, Institutul de Biologie București ș.a. Activitatea domniei sale a inclus și îndrumarea
cercului științific studențesc de botanică, în cadrul căruia s-au redactat anual teme, unele dintre
acestea fiind premiate la faza pe universitate a Sesiunii de comunicări științifice studențești sau chiar
la faza pe țară.
Retras din activitate, cu probleme de sănătate, de care însă nu s-a lăsat niciodată copleșit, tot
mai îngrijorat de starea fizică fragilă a doamnei, care din păcate a trecut în neființă în anul 2022,
domnul Profesor Danciu își reîncărca bateriile cu ocazia vizitelor fiului său Daniel și a celor 2 nepoți
- un nepoțel și o nepoțică, de care era foarte mândru și ori de câte ori, vizitându-1, ne arăta poze cu
ei, iar sufletul i se umplea de bucurie și ochii i se umezeau.
Zilele trecute, când ar fi trebuit să-l sărbătorim la vârsta de 82 de ani, vârstă pe care din păcate
nu a mai împlinit-o, nu am putut decât să ni-1 amintim așa cum l-am cunoscut, pentru că dumnealui
ne-a părăsit, trecând în eternitate, pe data de 26 aprilie 2023, la 81 de ani, 8 luni și 10 zile.
Chiar dacă problemele din ultimii ani l-au determinat să vină mai rar pe la facultate, mentorul
nostru, domnul Profesor dr. biol. Marius Danciu va rămâne, prin tot ceea ce a realizat, prin tot ceea
ce am învățat de la dumnealui, pentru totdeauna în amintirea noastră, a disciplinei de Botanică și
Fiziologia plantelor, a Facultății de Silvicultură și exploatări forestiere de la Brașov și a numeroase
generații de absolvenți ai acesteia.
Vă mulțumim Domnule Profesor și vă asigurăm de întreaga noastră gratitudine și
recunoștință ! Dumnezeu să vă odihnească !
44