Development, evaluation and application of a novel markerless motion analysis system to understand push-start technique in elite skeleton athletes

Abstract

This study describes the development, evaluation and application of a computer vision and deep learning system capable of capturing sprinting and skeleton push start step characteristics and mass centre velocities (sled and athlete). Movement data were captured concurrently by a marker-based motion capture system and a custom markerless system. High levels of agreement were found between systems, particularly for spatial based variables (step length error 0.001 ± 0.012 m) while errors for temporal variables (ground contact time and flight time) were on average within ± 1.5 frames of the criterion measures. Comparisons of sprinting and pushing revealed decreased mass centre velocities as a result of pushing the sled but step characteristics were comparable to sprinting when aligned as a function of step velocity. There were large asymmetries between the inside and outside leg during pushing (e.g. 0.22 m mean step length asymmetry) which were not present during sprinting (0.01 m step length asymmetry). The observed asymmetries suggested that force production capabilities during ground contact were compromised for the outside leg. The computer vision based methods tested in this research provide a viable alternative to marker-based motion capture systems. Furthermore, they can be deployed into challenging, real world environments to non-invasively capture data where traditional approaches are infeasible.

Automating Video‐Based Two‐Dimensional Motion Analysis in Sport? Implications for Gait Event Detection, Pose Estimation, and Performance Parameter Analysis

“…The FCA [24] relies on a TOE marker to detect TO. Using the 2D projected keypoints, TO was detected with a mean bias of 44 and 43 ms for frame rates of 200 and 100 Hz (SD: 100 and 97 ms, LoA: −152 to 240 and −147 to 233 ms), which is in a similar range to TD detection.…”

“…The use of 2D OpenPose and projected keypoint data as inputs resulted in an overall loss of accuracy and precision. 2D projected keypoints showed a mean bias in TD detection of 38 The FCA [24] relies on a TOE marker to detect TO. Using the 2D projected keypoints, TO was detected with a mean bias of 44 and 43 ms for frame rates of 200 and 100 Hz (SD: 100 and 97 ms, LoA: −152 to 240 and −147 to 233 ms), which is in a similar range to TD detection.…”

“…Currently, there is no evidence that 3D gait event detection algorithms can be adapted for use with 2D PEMS, or that using 2D keypoints in place of 3D marker trajectories is a valid substitution. Other computer vision‐based algorithms have been proposed for detecting key events in running, but the bulk of these rely on multiple cameras [23, 24], require computationally high‐cost and low‐accuracy background subtraction techniques [25], or suffer from low‐accuracy motion blur [26].…”

Ubiquitous Computing in Sports and Physical Activity—Recent Trends and Developments

“…However, movement velocity and foot–ball contact might increase the errors associated with the estimated kinematics. Needham et al [ 24 ] proposed and assessed a computer vision method based on pose estimation to provide accurate estimates of body centre of mass (CoM) kinematics of athletes and the sled during skeleton bob races. MoCap and OpenPose data capture were synchronised for assessment and comparison of sprint start and skeleton push start stride characteristics and mass centre velocities for both athlete and sled.…”