Keynote Title: Can Markerless Tracking of Human Motion Revolutionize the Analysis of In-game Sports Performance?
Keynote Lecturer: Scott Selbie
Presented on: 05/11/2020, Online Streaming
Abstract: For some time now, researchers in biomechanics have envisioned a new paradigm where accurate data about human movement would provide a basis for analyzing in-game performance. This presentation is predicated on the proposition that 3D biomechanical analyses of player performance based on in-game recordings can be used to evaluate performance, predict the susceptibility to musculoskeletal injuries, inform training and rehabilitation strategies, and influence game strategy. If this proposition is true, markerless tracking of human motion will revolutionize the analysis of in-game sports performance. As this talk is about the performance of players on the field I will summarize the motion tracking of players that is possible during live game recording, and will focus on 3D human motion tracking using temporally and spatially synchronized commercial video cameras mounted around the field of play. Video based solutions passively observe the field of play, can be used indoors and outdoors, and do not require that sensors or markers be attached to the players; thus do not require the player’s or team’s awareness of the recording, cooperation, or coordination, and do not influence the game directly.I will present the state of the art 3D tracking of human motion suitable for biomechanical analyses based on using Deep Similarity Learning to identify uniquely all players in the field of view of the cameras, Deep Neural Networks to identify the location of anatomical features of each player in the multiple video images, and multibody optimization to consolidate this data into a mathematically observable estimate of the 3D position and orientation (pose) of a personalized model of each player. Such systems do not require assumptions of symmetry, behavioural rules or physical constraints as they directly estimate the 3D pose. The nature of the technology lends itself to the vast amounts of accurate and meaningfully consolidated data required by the burgeoning field of sports analytics. I will present experimental results of the critical issues of accuracy, repeatability and reliability of the pose estimation in a controlled laboratory setting and speculate on these same issues in live game recording.The potential for a biomechanical analysis of individual player performance is actual. As an example of the scope of data that is now possible, at the time of submission of this abstract Kinatrax Inc (USA) had recorded and analyzed more than 560,000 pitches during live in season Major League Baseball games.
Presented at the following Conference: icSPORTS, 8th International Conference on Sport Sciences Research and Technology Support