Race Walking Computer Vision - Determining Legality of an Athlete

Computer Vision application that will detect and analyze the precise movement phases in Race Walking. With the help of AWS (SageMaker, S3, Lambda), OpenShot, and TensorFlow, our team is able to capture and label specific frames from video footage to train a model that will identify foot-ground contact points. These phases include heel strikes, full foot contact, toe contact, and toe-off, as well as scenarios when the foot is not touching the ground. The trained model will help identify the correct form and foot-ground contact in Race Walking, this will help athletes and coaches receive accurate feedback and analysis better.