Accuracy of image data stream of a markerless motion capture system in determining the local dynamic stability and joint kinematics of human gait

Correlation between MOVA3D, a Monocular Movement Analysis System, and Qualisys Track Manager (QTM) during Lower Limb Movements in Healthy Adults: A Preliminary Study

New technologies based on virtual reality and augmented reality offer promising perspectives in an attempt to increase the assessment of human kinematics. The aim of this work was to develop a markerless 3D motion analysis capture system (MOVA3D) and to test it versus Qualisys Track Manager (QTM). A digital camera was used to capture the data, and proprietary software capable of automatically inferring the joint centers in 3D and performing the angular kinematic calculations of interest was developed for such analysis. In the experiment, 10 subjects (22 to 50 years old), 5 men and 5 women, with a body mass index between 18.5 and 29.9 kg/m2, performed squatting, hip flexion, and abduction movements, and both systems measured the hip abduction/adduction angle and hip flexion/extension, simultaneously. The mean value of the difference between the QTM system and the MOVA3D system for all frames for each joint angle was analyzed with Pearson's correlation coefficient (r). The MOVA3D system reached good (above 0.75) or excellent (above 0.90) correlations in 6 out of 8 variables. The average error remained below 12° in only 20 out of 24 variables analyzed. The MOVA3D system is therefore promising for use in telerehabilitation or other applications where this level of error is acceptable. Future studies should continue to validate the MOVA3D as updated versions of their software are developed.

Motion Capture Technologies for Ergonomics: A Systematic Literature Review

Muscular skeletal disorder is a difficult challenge faced by the working population. Motion capture (MoCap) is used for recording the movement of people for clinical, ergonomic and rehabilitation solutions. However, knowledge barriers about these MoCap systems have made them difficult to use for many people. Despite this, no state-of-the-art literature review on MoCap systems for human clinical, rehabilitation and ergonomic analysis has been conducted. A medical diagnosis using AI applies machine learning algorithms and motion capture technologies to analyze patient data, enhancing diagnostic accuracy, enabling early disease detection and facilitating personalized treatment plans. It revolutionizes healthcare by harnessing the power of data-driven insights for improved patient outcomes and efficient clinical decision-making. The current review aimed to investigate: (i) the most used MoCap systems for clinical use, ergonomics and rehabilitation, (ii) their application and (iii) the target population. We used preferred reporting items for systematic reviews and meta-analysis guidelines for the review. Google Scholar, PubMed, Scopus and Web of Science were used to search for relevant published articles. The articles obtained were scrutinized by reading the abstracts and titles to determine their inclusion eligibility. Accordingly, articles with insufficient or irrelevant information were excluded from the screening. The search included studies published between 2013 and 2023 (including additional criteria). A total of 40 articles were eligible for review. The selected articles were further categorized in terms of the types of MoCap used, their application and the domain of the experiments. This review will serve as a guide for researchers and organizational management.

Walking Speed Classification from Marker-Free Video Images in Two-Dimension Using Optimum Data and a Deep Learning Method

Walking speed is considered a reliable assessment tool for any movement-related functional activities of an individual (i.e., patients and healthy controls) by caregivers and clinicians. Traditional video surveillance gait monitoring in clinics and aged care homes may employ modern artificial intelligence techniques to utilize walking speed as a screening indicator of various physical outcomes or accidents in individuals. Specifically, ratio-based body measurements of walking individuals are extracted from marker-free and two-dimensional video images to create a walk pattern suitable for walking speed classification using deep learning based artificial intelligence techniques. However, the development of successful and highly predictive deep learning architecture depends on the optimal use of extracted data because redundant data may overburden the deep learning architecture and hinder the classification performance. The aim of this study was to investigate the optimal combination of ratio-based body measurements needed for presenting potential information to define and predict a walk pattern in terms of speed with high classification accuracy using a deep learning-based walking speed classification model. To this end, the performance of different combinations of five ratio-based body measurements was evaluated through a correlation analysis and a deep learning-based walking speed classification test. The results show that a combination of three ratio-based body measurements can potentially define and predict a walk pattern in terms of speed with classification accuracies greater than 92% using a bidirectional long short-term memory deep learning method.

A SWOT Analysis of Portable and Low-Cost Markerless Motion Capture Systems to Assess Lower-Limb Musculoskeletal Kinematics in Sport

Markerless motion capture systems are promising for the assessment of movement in more real world research and clinical settings. While the technology has come a long way in the last 20 years, it is important for researchers and clinicians to understand the capacities and considerations for implementing these types of systems. The current review provides a SWOT (Strengths, Weaknesses, Opportunities, and Threats) analysis related to the successful adoption of markerless motion capture technology for the assessment of lower-limb musculoskeletal kinematics in sport medicine and performance settings. 31 articles met the a priori inclusion criteria of this analysis. Findings from the analysis indicate that the improving accuracy of these systems via the refinement of machine learning algorithms, combined with their cost efficacy and the enhanced ecological validity outweighs the current weaknesses and threats. Further, the analysis makes clear that there is a need for multidisciplinary collaboration between sport scientists and computer vision scientists to develop accurate clinical and research applications that are specific to sport. While work remains to be done for broad application, markerless motion capture technology is currently on a positive trajectory and the data from this analysis provide an efficient roadmap toward widespread adoption.

Motion envelopes: unfolding longitudinal rotation data from walking stick-figures

This work presents Motion Envelopes (ME), a simple method to estimate the missing longitudinal rotations of minimal stick figures, which is based on the spatial-temporal surface traced by line segments that connect contiguous pairs of joints. We validate ME by analyzing the gait patterns of 6 healthy subjects, comprising a total of 18 gait cycles. A strong correlation between experimental and estimated data was obtained for lower limbs and upper arms, indicating that ME can predict their longitudinal orientation in normal gait, hence, ME can be used to complement the kinematic information of stick figures whenever it is incomplete.

Sparse Visual-Inertial Measurement Units Placement for Gait Kinematics Assessment

This study investigates the possibility of estimating lower-limb joint kinematics and meaningful performance indexes for physiotherapists, during gait on a treadmill based on data collected from a sparse placement of new Visual Inertial Measurement Units (VIMU) and the use of an Extended Kalman Filter (EKF). The proposed EKF takes advantage of the biomechanics of the human body and of the investigated task to reduce sensor inaccuracies. Two state-vector formulations, one based on the use of constant acceleration model and one based on Fourier series, and the tuning of their corresponding parameters were analyzed. The constant acceleration model, due to its inherent inconsistency for human motion, required a cumbersome optimisation process and needed the a-priori knowledge of reference joint trajectories for EKF parameters tuning. On the other hand, the Fourier series formulation could be used without a specific parameters tuning process. In both cases, the average root mean square difference and correlation coefficient between the estimated joint angles and those reconstructed with a reference stereophotogrammetric system was 3.5deg and 0.70, respectively. Moreover, the stride lengths were estimated with a normalized root mean square difference inferior to 2% when using the forward kinematics model receiving as input the estimated joint angles. The popular gait deviation index was also estimated and showed similar results very close to 100, using both the proposed method and the reference stereophotogrammetric system. Such consistency was obtained using only three wireless and affordable VIMU located at the pelvis and both heels and tracked using two affordable RGB cameras. Being further easy-to-use and suitable for applications taking place outside of the laboratory, the proposed method thus represents a good compromise between accurate reference stereophotogrammetric systems and markerless ones for which accuracy is still under debate.

Local dynamic stability of the lower-limb as a means of post-hoc injury classification

Since most sporting injuries occur at the lower extremity (50% to 66%) and many of those injuries occur at the knee (30% to 45%), it is important to have robust metrics to measure risk of knee injury. Dynamic measures of knee stability are not commonly used in existing metrics but could provide important context to knee health and improve injury screening effectiveness. This study used the Local Dynamic Stability (LDS) of knee kinematics during a repetitive vertical jump to perform a post-hoc previous injury classification of participants. This study analyzed the kinematics from twenty-seven female collegiate division 1 (D1) soccer, D1 basketball, and club soccer athletes from Auburn University (height = 171 ± 8.9cm, weight = 66.3 ± 8.6kg, age = 19.8 ± 1.9yr), with 7 subjects having sustained previous knee injury requiring surgery and 20 subjects with no history of injury. This study showed that LDS correctly identified 84% of previously injured and uninjured subjects using a multivariate logistic regression during a fatigue jump task. Findings showed no statistical difference in kinematic position at maximum knee flexion during all jumps between previously injured and uninjured subjects. Additionally, kinematic positioning at maximum knee flexion was not indicative of LDS values, which would indicate that future studies should look specifically at LDS with respect to injury prevention as it cannot be effectively inferred from kinematics. These points suggest that the LDS preserves information about subtle changes in movement patterns that traditional screening methods do not, and this information could allow for more effective injury screening tests in the future.

Using a Deep Learning Method and Data from Two-Dimensional (2D) Marker-Less Video-Based Images for Walking Speed Classification

Human body measurement data related to walking can characterize functional movement and thereby become an important tool for health assessment. Single-camera-captured two-dimensional (2D) image sequences of marker-less walking individuals might be a simple approach for estimating human body measurement data which could be used in walking speed-related health assessment. Conventional body measurement data of 2D images are dependent on body-worn garments (used as segmental markers) and are susceptible to changes in the distance between the participant and camera in indoor and outdoor settings. In this study, we propose five ratio-based body measurement data that can be extracted from 2D images and can be used to classify three walking speeds (i.e., slow, normal, and fast) using a deep learning-based bidirectional long short-term memory classification model. The results showed that average classification accuracies of 88.08% and 79.18% could be achieved in indoor and outdoor environments, respectively. Additionally, the proposed ratio-based body measurement data are independent of body-worn garments and not susceptible to changes in the distance between the walking individual and camera. As a simple but efficient technique, the proposed walking speed classification has great potential to be employed in clinics and aged care homes.

Automatic Diagnosis of Cerebral Palsy Gait Using Computational Intelligence Techniques: A Low-Cost Multi-Sensor Approach

Automatic diagnosing of Cerebral Palsy (CP) gait is crucial in quantitative evaluation of a therapeutic intervention. Existing systems for such gait assessment are expensive and require user intervention. This study proposes a low-cost gait assessment system equipped with multiple Kinect sensors. Forty subjects (20 CP patients and 20 normal) were recruited for the experiment. To remove outlier frames from the combined gait signal of multiple sensors a data driven algorithm was proposed. Different supervised classifiers along with extreme learning machine were investigated to diagnose CP gait. In addition, a feature level analysis was also performed. Several spatio-temporal features (i.e. step length, stride length, stride time, etc.) were extracted. The strength of walking ratio, a speed invariant feature, to detect CP gait was thoroughly analyzed. The proposed system outperformed state-of-the-art with ≈98% of accuracy (sensitivity: 100%, and specificity: 96.87%). Results indicate a substantial improvement in abnormality detection performance after outlier removal. Based on ReliefF feature ranking algorithm, walking ratio ranked the best among other classical gait features. Performance of all classifiers increased substantially using walking ratio as a feature. Extreme learning machine demonstrated a competing performance in all cases. The higher classification accuracy of this low-cost system using only a single feature makes it attractive for CP gait detection.