Concurrent assessment of gait kinematics using marker-based and markerless motion capture

Abstract

Kinematic analysis is a useful and widespread tool used in research and clinical biomechanics for the estimation of human pose and the quantification of human movement. Common marker-based optical motion capture systems are expensive, time intensive, and require highly trained operators to obtain kinematic data. Markerless motion capture systems offer an alternative method for the measurement of kinematic data with several practical benefits. This work compared the kinematics of human gait measured using a deep learning algorithm-based markerless motion capture system to those of a common marker-based motion capture system. Thirty healthy adult participants walked on a treadmill while data were simultaneously recorded using eight video cameras (markerless) and seven infrared optical motion capture cameras (marker-based). Video data were processed using markerless motion capture software, marker-based data were processed using marker-based capture software, and both sets of data were compared. The average root mean square distance (RMSD) between corresponding joints was less than 3 cm for all joints except the hip, which was 4.1 cm. Lower limb segment angles indicated pose estimates from both systems were very similar, with RMSD of less than 6° for all segment angles except those that represent rotations about the long axis of the segment. Lower limb joint angles captured similar patterns for flexion/extension at all joints, ab/adduction at the knee and hip, and toe-in/toe-out at the ankle. These findings demonstrate markerless motion capture can measure similar 3D kinematics to those from marker-based systems.

Assessment of spatiotemporal gait parameters using a deep learning algorithm-based markerless motion capture system

“…Theia3D (Theia Markerless Inc., Kingston, ON, Canada) is a deep learning algorithm-based approach to markerless motion capture which uses deep convolutional neural networks for feature recognition (humans and human features) within 2D camera views (Kanko et al, 2020a(Kanko et al, , 2020bMathis and Mathis, 2020). The neural networks were trained on digital images of over 500,000 humans in the wild.…”

“…By default, the lower body kinematic chain has six degrees-of-freedom (DOF) at the pelvis, and three DOF at the hip, knee, and ankle. This markerless system has been shown to measure gait kinematics similarly to marker-based systems, and with greater reliability over multiple sessions (Kanko et al, 2020a(Kanko et al, , 2020b.…”

Pose2Sim: An End-to-End Workflow for 3D Markerless Sports Kinematics—Part 1: Robustness

“…They compared their results to those of IMU and pointed out that errors were higher for running than for walking, and were also rather inconsistent (by up to 14°), although they decreased from 2° to 7° if the two cameras were set laterally rather than at the back of the subject. Theia3D, a recent commercial (and not open) solution, estimates the positions of a set of keypoints around the joint and then uses a multi-body optimization approach to solve inverse kinematics (See Kanko et al [ 56 , 57 ]). They noticed an offset in hip and ankle angles between their markerless system and the reference marker-based one, likely due to different skeletal models.…”

Biomechanical Analysis of the Throwing Athlete and Its Impact on Return to Sport

“…A possible new frontier exists in marker-less tracking. Studies have begun to apply this technology for analysis of gait, [50][51][52] jumping, 53 and baseball swinging. 21 Application of marker-less technology to throwing has not been sufficiently validated to date, but research is ongoing, In a study comparing marker-based to marker-less motion capture during gait, differences between the 2 techniques were only 2.1 cm, 2.4 cm, and 1.1 cm for segment locations at the shoulder, elbow, and wrist, respectively.…”

“…21 Application of marker-less technology to throwing has not been sufficiently validated to date, but research is ongoing, In a study comparing marker-based to marker-less motion capture during gait, differences between the 2 techniques were only 2.1 cm, 2.4 cm, and 1.1 cm for segment locations at the shoulder, elbow, and wrist, respectively. 50 In 2020, Major League Baseball also introduced a motion capture system for in-game use. This system (Hawk-Eye Statcast; Hawk-Eye Innovations Ltd., Basingstoke, UK) relies on 12 cameras positioned around an MLB ballpark.…”