Sport Data

Sports data analysis leverages computer vision and machine learning to extract meaningful insights from video and sensor data, aiming to improve performance analysis, training strategies, and fan engagement. Current research focuses on developing robust algorithms for tasks such as player tracking (using models like HM-SORT and SportsTrack), pose estimation (AutoSoccerPose), and activity recognition (NETS), often employing transformer-based architectures and deep learning techniques. These advancements are significantly impacting sports analytics, enabling more objective performance evaluations, automated highlight generation, and data-driven decision-making for coaches and athletes.