Degree level: Masters CS or AI
Supervisor: Mauricio Verano Merino, Idsart Kingma (Faculty of Behavioural and Movement Sciences)
Problem description
The research project aims to develop a system for goalkeepers that transforms 2D key point detections from video footage into accurate 3D representations. Using computer vision techniques, the system will detect key points on the goalkeeper’s body (such as joints, limbs, and hands) from multiple camera angles or a monocular setup. These 2D key points will then be used to reconstruct a 3D model of the goalkeeper’s posture and movements in real-time. The project seeks to improve goalkeeper training by providing precise, in-depth analysis of their positioning, reactions, and decision-making during matches and drills. This could also enhance biomechanical feedback and inform advanced motion prediction models.