Are we there yet? A systematic review and meta-analysis of the validity and reliability of automated markerless motion capture systems during jumping tasks

Dynamic assessment of spine movement patterns using an RGB-D camera and deep learning

Basketball requires high lower limb performance. Assessing jump biomechanics is vital for enhancing performance and injury prevention. Marker-based (MB) systems are common but limited. In recent years, Markerless (ML) motion capture systems have gradually become emerging tools in sports biomechanics research due to their characteristic of not requiring physical marker points. However, their specific application and verification in basketball events are still relatively limited. Purpose: In this study, lower limb kinematics and kinetics estimated by MB and ML motion capture systems during jumps were compared. Methods: Twelve subjects performed the standing vertical jump (SVJ), standing long jump (SLJ) and running vertical jump (RVJ) tests. Data was collected using 10 infrared cameras, 6 high-resolution cameras and two force platforms via Vicon Nexus software. Markerless motion capture calculated sagittal plane angles, torque and power of the Hip, Knee and Ankle joints via Theia3D software, with these parameters also collected by the marker-based Vicon system. Both systems' '64ata were then processed in Visual3D. We analyzed the correlation coefficient (r), root mean square difference (RMSD), and maximum/minimum errors, as well as using statistical parametric mapping (SPM) to compare temporal patterns between groups and determine specific moments where significant differences occurred. Results: SLJ capture was slightly inferior in both systems. SPM analysis of the sagittal plane showed significant differences only at the hip joint. Joint angle RMSD was < 8.2°, torque RMSD < 0.41 N·M/kg, and power RMSD < 1.76 W/kg. Conclusions: The ML system accurately captures knee and ankle joints in the sagittal plane but shows significant differences in hip measurement and certain movements, requiring further validation.

Validity and Reliability of Upper Limb Kinematic Assessment Using a Markerless Motion Capture (MMC) System: A Pilot Study

BackgroundInvestigations into the possible associations between early in life motor function and later in life musculoskeletal health, will require easily obtainable, valid, and reliable measures of gross motor function and kinematics. Marker-based motion capture systems provide reasonably valid and reliable measures, but recordings are restricted to expensive lab environments. Markerless motion capture systems can provide measures of gross motor function and kinematics outside of lab environments and with minimal interference to the subjects being investigated. It is, however, unknown if these measures are sufficiently valid and reliable in young children to warrant further use. This study aims to document the concurrent validity of a markerless motion capture system: “The Captury.”MethodMeasures of gross motor function and lower extremity kinematics from 14 preschool children (age between three and 6 years) performing a series of squats and standing broad jumps were recorded by a marker-based (Vicon) and a markerless (The Captury) motion capture system simultaneously, in December 2015. Measurement differences between the two systems were examined for the following variables: jump length, jump height, hip flexion, knee flexion, ankle dorsi flexion, knee varus, knee to hip separation distance ratio (KHR), ankle to hip separation distance ratio (AHR), frontal plane projection angle, frontal plane knee angle (FPKA), and frontal plane knee deviation (FPKD). Measurement differences between the systems were expressed in terms of root mean square errors, mean differences, limits of agreement (LOA), and intraclass correlations of absolute agreement (ICC (2,1) A) and consistency of agreement.ResultsMeasurement differences between the two systems varied depending on the variables. Agreement and reliability ranged from acceptable for e.g. jump height [LOA: − 3.8 cm to 2.2 cm; ICC (2,1) A: 0.91] to unacceptable for knee varus [LOA: − 33° to 19°; ICC (2,1) A: 0.29].ConclusionsThe measurements by the markerless motion capture system “The Captury” cannot be considered interchangeable with the Vicon measures, but our results suggest that this system can produce estimates of jump length, jump height, KHR, AHR, knee flexion, FPKA, and FPKD, with acceptable levels of agreement and reliability. These variables are promising for use in future research but require further investigation of their clinimetric properties.

Markerless (ML) motion capture has emerged as a viable option to marker-based (MB) motion capture in estimating movement biomechanics, but limited data exists on the accuracy of ML systems during high-speed throwing. This study evaluated the accuracy and reliability of an in-stadium (Hawk-Eye) and a portable (Theia3D) ML motion-capture system in quantifying baseball pitching kinematics and kinetics relative to an MB reference. Eighteen collegiate pitchers were simultaneously recorded using all three systems. Mean per-joint position error (MPJPE), statistical parametric mapping (SPM), root mean square error (RMSE), Bland-Altman analysis, and concordance correlation coefficients (CCC) were used to assess agreement. Both ML systems demonstrated measurable discrepancies across variables, with MPJPE values of 56.6 ± 9.4 mm (Hawk-Eye) and 52.0 ± 12.3 mm (Theia3D). Stride length exhibited the strongest agreement with MB in both systems (CCC > 0.85), whereas shoulder rotational variables showed greater variability. Error magnitudes in joint positions and kinematic waveforms were comparable to those reported for other ML systems during dynamic movements. These results highlight the influence of system configuration, camera deployment, and pose-estimation models on biomechanical accuracy. Overall, both configurations showed potential for estimating pitching biomechanics, underscoring the trade-offs between criterion and ecological validity in markerless motion capture.

Markerless motion capture (MMC) is a newer technology that utilizes advanced computer vision and machine learning algorithms to evaluate human movement. Traditional marker-based motion capture (MBC) systems can be time consuming for both participant and researcher and setup can often take more time than collection. Further, MMC allows athletes to perform in a more familiar environment without the potential error introduced by the laboratory environment. However, it is imperative to examine its accuracy in comparison to the gold-standard of conventional MBC. PURPOSE: To determine interrater reliability between a marker-based and markerless motion capture systems. METHODS: 15 collegiate basketball players (8 females, height: 180.8 ± 14.0 cm, mass: 80.8 ± 19.0 kg) participated in this study. 3 double leg drop vertical jumps (DVJ) and 3 maximum single leg vertical countermovement hops on each limb were performed. Limb dominance (Dom vs NDom) was self-reported through survey. Standard MBC techniques were utilized to track markers throughout the execution of each task. A MMC system collected data concurrently and used advanced computer vision and machine learning algorithms to estimate the pose of individuals during dynamic movements without the need of markers. Hip joint centers were estimated from each system and used to determine the lowest and highest vertical displacement during each trial. Intraclass correlation coefficient (ICC[3,1]) was used to determine the interrater reliability between systems. ICC and 95% confidence interval were used for interpretation. RESULTS: Interrater reliability between MMC and MBC was excellent for DOM side single leg jumps (MMC 47.8 ± 10.6 cm, MBC 47.1 ± 10.8 cm, ICC = 0.986 [95 CI 0.966, 0.994]), NDom side single leg jumps (MMC 48.4 ± 10.3 cm, MBC 47.2 ± 10.7 cm, ICC = 0.983 [95 CI 0.958, 0.993]), and double leg drop vertical jumps (MMC 67.7 ± 11.1 cm, MBC 66.9 ± 12.1 cm, ICC = 0.982 [95 CI 0.956, 0.993]). The root mean square error between systems was less than 2.3 cm (Dom: 1.8 cm, NDom: 2.2 cm, DVJ: 2.3 cm). CONCLUSIONS: Markerless motion capture techniques to calculate vertical jump displacement had excellent agreement to marker-based motion capture during both single leg and double leg tasks. Future studies should compare additional kinematic and kinetic variables between systems.

IntroductionAdvances in motion capture technology include markerless systems to facilitate valid data collection. Recently, the technological reliability of this technology has been reported for human movement assessments. To further understand sources of potential error, biological reliability must also be determined. The aim of this study was to determine the day-to-day reliability for a three-dimensional markerless motion capture (MMC) system to quantify 4 movement analysis composite scores, and 81 kinematic variables.MethodsTwenty-two healthy men (n = 11; ; age = 23.0 ± 2.6 years, height = 180.4.8 cm, weight = 80.4 ± 7.3 kg) and women (n = 11; age = 20.8 ± 1.1 years, height = 172.2 ± 7.4 cm, weight = 68.0 ± 7.3 kg) participated in this study. All subjects performed 4 standardized test batteries consisting of 14 different movements on four separate days. A three-dimensional MMC system (DARI Motion, Lenexa, KS) using 8 cameras surrounding the testing area was used to quantify movement characteristics. 1 × 4 RMANOVAs were used to determine significant differences across days for the composite movement analysis scores, and RM-MANOVAs were used to determine test day differences for the kinematic data (p < 0.05). Intraclass correlation coefficients (ICCs) were reported for all variables to determine test reliability. To determine biological variability, mean absolute differences from previously reported technological variability data were subtracted from the total variability data from the present study.ResultsNo differences were observed for any composite score (i.e., athleticism, explosiveness, quality, readiness; or any of the 81 kinematic variables. Furthermore, 84 of 85 measured variables exhibited good to excellent ICCs (0.61–0.99). When compared to previously reported technological variability data, 62.3% of item variability was due to biological variability, with 66 of 85 variables exhibiting biological variability as the primary source of error (i.e., >50% total variability).DiscussionCombined, these findings effectively add to the body of literature suggesting sufficient reliability for MMC solutions in capturing kinematic features of human movement.

Background:The range-of motion (ROM) is an essential component of joint mobility. Shoulder ROM measurement has been problematic due to its complexity. A marker less motion capture system can be a potential alternative for upper limb assessment. Currently, there is no systematic review to evaluate the validity of a marker less motion capture system for assessing shoulder ROM. This study aims to describe methods to evaluate the reliability and validity of a single camera marker less motion capture system that uses an RGB-depth sensor to measure shoulder ROM.Methods:Studies that measured shoulder ROM with a single camera marker less motion capture system using the RGB-depth sensor and assessed the intra- and/or inter-rater reliability, and/or validity of the device will be included. The search of electronic databases, such as MEDLINE, EMBASE, Cochran library, Cumulative Index to Nursing, and Allied Health Literature via EBSCO, IEEE Xplore, China National Knowledge Infrastructure, KoreaMed, Korean studies Information Service System, and Research Information Sharing Services will be performed for all relevant articles from inception to December 2022. Two authors will independently perform quality assessments using the Consensus-based Standards for the selection of health Measurement Instruments checklist for reliability, measurement error of outcome measurement instrument, and criterion validity. The primary outcomes will be the intra- and inter-rater reliability and validity of the markerless motion capture system measuring shoulder flexion, extension, abduction, adduction, internal rotation, or external rotation. A subgroup analysis would be performed if there are sufficient data to pool to identify an influencing factor in the measurement of ROM using a marker less motion capture system.Results and Conclusion:These findings will present tools to utilize and evaluate single camera motion capture systems for the medical use for clinicians and healthcare experts and can aid in further clinical research using such a system for different movements and other joints.

Background: Markerless (ML) motion capture systems have recently become available for biomechanics applications. Evidence has indicated the potential feasibility of using an ML system to analyze lower extremity kinematics. However, no research has examined ML systems’ estimation of the lower extremity joint moments and powers. This study aimed to compare lower extremity joint moments and powers estimated by marker-based (MB) and ML motion capture systems. Methods: Sixteen volunteers ran on a treadmill for 120 s at 3.58 m/s. The kinematic data were simultaneously recorded by 8 infrared cameras and 8 high-resolution video cameras. The force data were recorded via an instrumented treadmill. Results: Greater peak magnitudes for hip extension and flexion moments, knee flexion moment, and ankle plantarflexion moment, along with their joint powers, were observed in the ML system compared to an MB system (p < 0.0001). For example, greater hip extension (MB: 1.42 ± 0.29 vs. ML: 2.27 ± 0.45) and knee flexion (MB: −0.74 vs. ML: −1.17 nm/kg) moments were observed in the late swing phase. Additionally, the ML system’s estimations resulted in significantly smaller peak magnitudes for knee extension moment, along with the knee production power (p < 0.0001). Conclusions: These observations indicate that inconsistent estimates of joint center position and segment center of mass between the two systems may cause differences in the lower extremity joint moments and powers. However, with the progression of pose estimation in the markerless system, future applications can be promising.

Instrumented gait analysis has traditionally been isolated to laboratory, marker-based optoelectronic motion capture systems, which limits clinical uptake. Markerless motion capture (MMC) systems driven by trained machine learning algorithms offer high-throughput solutions for translational clinical opportunities. The aim of this study was to examine the day-to-day repeatability of discrete knee kinematic gait metrics in a healthy population using an MMC system uniquely installed in a hospital hallway. Twenty healthy adults (13 females, 7 males) participated in 3 overground hallway gait sessions, on average 11days apart, using a novel MMC system setup. Intraclass correlation coefficients, standard errors of measurement, and minimal detectable changes were examined for each gait outcome. Results indicated good-to-excellent repeatability, with most (7/8) outcomes having intraclass correlation coefficient values over .86. Standard error of measurement values for all kinematic outcomes were less than 2.0°, and minimal detectable change values were less than 4.7°. Our novel setup of a hospital hallway MMC system produced highly repeatable gait kinematic metrics in a population of healthy adults. Repeatability errors from this study can be used as a healthy reference for future applications of this system.

Comparison of kinematics and kinetics between OpenCap and a marker-based motion capture system in cycling.

Motion capture technology is quickly evolving, providing researchers, clinicians, and coaches with more access to biomechanics data. Markerless motion capture and inertial measurement units (IMUs) are continually developing biomechanics tools that need validation for dynamic movements before widespread use in applied settings. This study evaluated the validity of a markerless motion capture, IMU, and red, green, blue, and depth (RGBD) camera system as compared with marker-based motion capture during countermovement jumps, overhead squats, lunges, and runs with cuts. Thirty adults were recruited for this study (sex: 18 females, 12 males; age: 25.4 ± 8.6 yrs; height: 1.71 ± 0.08 m; weight: 71.6 ± 11.5 kg). Data were collected simultaneously with four motion capture technologies (i.e., Vicon, marker-based; Theia/Optitrack, markerless; APDM Opals, IMUs; and Vald HumanTrak, RGBD camera). System validity for lower and upper body joint angles was evaluated using bias, root mean squared error (RMSE), precision, maximum absolute error, and intraclass correlation coefficients. System usability was descriptively analyzed. Overall, markerless motion capture had the highest validity (sagittal plane RMSE: 3.20°-15.66°; frontal plane RMSE: 2.12°-9.14°; transverse plane RMSE: 3.160°-56.61°), followed by the IMU system (sagittal plane RMSE: 8.11°-28.37°; frontal plane RMSE: 3.26°-16.98°; transverse plane RMSE: 5.08°-116.75°), and lastly the RGBD system (sagittal plane bias: 0.55°-129.48°; frontal plane bias: 1.35°-52.06°). Markerless motion capture and IMUs have moderate validity for joint kinematics, whereas the RGBD system did not have adequate validity. Markerless systems have lower data processing time, require moderate technical expertise, but have high data storage size. IMUs are easier to use, can collect data in any location, but require participant set-up. Overall, individuals using motion capture should consider the specific movements, testing locations, and technical expertise available before selecting a system.

BackgroundCommercial markerless motion capture (MMC) systems show promise for use in rehabilitation and have been validated for the assessment of various parameters. However, no prior studies have evaluated MMC systems to detect stance asymmetry.ObjectiveThe objective of this study was to assess the accuracy and variability of the Jintronix Weight Shift Tool MMC system to estimate the percentage of weight borne on each foot.MethodsTwelve healthy younger adults, 12 healthy older adults, and 12 people living with stroke were recruited for this cross-sectional study. The percentage of weight borne on each foot was simultaneously recorded by the Weight Shift Tool and a validated pressure mat during 2 series (raising the arm to capture the recording and without arm raise) with left lean, right lean, and equal stance. Agreement between the Weight Shift Tool and the pressure mat was assessed using Bland-Altman analyses.ResultsBias was greatest for older adults for all stances except for right lean without arm raise. Variability was greatest for people living with stroke for all stances except for left lean with arm raise. On average, the limits of agreement were narrower during equal stance. Although bias between the Weight Shift Tool and the pressure mat was small to moderate (0.0%-11.7%), the limits of agreement were wide (12.8%-33.6% above and below the bias).ConclusionsThe Weight Shift Tool is not clinically acceptable for the estimation of the percentage of weight on each foot due to high variability. Investigation of other MMC systems is required to confirm the validity of MMC for clinical assessment.

Markerless motion capture (MMC) shows promise for examining human movement across many domains because of its nonintrusive nature and negligible per-subject setup time. However, published MMC systems typically require specific hardware. This validation study compared lower-body joint kinematics from Ergomechanic, a hardware-agnostic pose model-based MMC system, to an established marker-based motion capture (MBMC) system. Static trial data from eighteen people were used to register MMC keypoints within a widely used musculoskeletal model. The registered model was used to calculate joint kinematics for static pose, squatting, and walking trials. A novel perturbation analysis estimated the contributions to differences in MBMC and MMC approaches to measurement disparities. Very good (0.87-1.0) correlations between the systems were calculated for ankle, knee, and hip flexion-extension angles. Good (0.70-0.86) correlations were found for hip external-internal and abduction-adduction. Pelvis and lumbar spine angles had a wider range of correlation results (-0.06 to 0.95), likely due to the few MMC keypoints in these body regions. Relative contributions from the perturbation analysis were (i) 75% from variations in MMC data relative to MBMC; (ii) 8% because MMC keypoints (26) < MBMC markers (67); and (iii) 3% from differences in musculoskeletal model scaling. These results validate Ergomechanic for leg kinematics during standing, walking, and squatting. Further, they suggest system improvements for pelvis and torso kinematics and provide new insights into the sources of known differences between MMC and MBMC measurements.

Agreement between a markerless and a marker-based motion capture systems for balance related quantities