The AI Coaching Assistant (ACA) project aims to create an athlete performance and analysis systems. The project is co-sponsored by Gymshark and Aston university. The project applies the rapidly improving computer vision (markerless-pose) 2D and 3D models as an alternative data generation source for creating accessible performance analysis tools with focus on Olympic Weightlifting, a technically complex sport, where traditionally the movement analysis has been left in the domain of experienced coaches. We aim to show that with the use of currently available computer vision solutions and models developed for Action Quality Assessment with the help of modern Feature attribution methods this analysis can be made accessible to newcomers to the sport and developing coaches.
This work is part of the PhD project of Gatis Jurkans.

References

autoWT A Semi-Automated ML-Based Movement Tracking System for Performance Tracking and Analysis in Olympic Weightlifting. Gatis Jurkans, Anikó Ekárt, and Ulysses Bernardet. In Andrea Aliverti, Pedro G. Morouço, and Carlo Capelli, editors, Proceedings of the 12th International Conference on Sport Sciences Research and Technology Support, icSPORTS 2024, Porto, Portugal, November 21-22, 2024, pages 60--71. SCITEPRESS, 2024.