Phase aware modeling improves throw distance prediction from poses
Throwing4: Phase-Aware Pose Modeling for Throw-Distance Prediction in Elite Athletics
Computer Vision and Pattern Recognition
Summary
Predicting how far an athlete throws an object like a javelin or discus can be tricky because the motion involves many stages. The authors created a dataset with videos and pose data from elite athletes and divided each throw into four phases. They found that when their computer models pay attention to these phases separately, they can predict the throw distance more accurately than looking at the whole throw as one sequence. This means knowing the different parts of the throw helps computers understand and measure athletic performance better.
What this means in practice
- •For sports analytics teams: Improve performance prediction accuracy by incorporating biomechanical phase information into models analyzing athlete throw poses from broadcast videos.
- •For broadcast video analysts: Use phase-aware pose models to extract more reliable throw distance estimates from standard sports broadcasts without specialized motion capture equipment.
Authors
Bikash Kumar Badatya, Kartike Tiwari, Jyotirmoy Amin, Ravi Hegde
Abstract
Elite athletics throwing provides a structured biomechanical action sequence together with an objective continuous outcome, namely throw distance, making it a suitable setting for performance-oriented sports analytics. In this work, we investigate whether explicit phase alignment of pose dynamics improves throw-distance prediction under realistic broadcast conditions. To support this study, we curate Throwing4, a dataset of 500 elite throwing clips spanning javelin, discus, hammer, and shot put, with corresponding two-dimensional pose sequences, frame-level phase annotations, and official distance labels. We formulate throw-distance prediction from broadcast-derived pose as a regression problem and compare full-sequence modeling against a phase-conditioned approach that partitions each attempt into four technique stages and fuses phase-specific motion representations. Experiments across multiple graph-based pose encoders show a consistent reduction in prediction error when phase structure is incorporated, indicating that phase-aligned modeling provides a more stable representation than treating the entire motion as an unstructured sequence. These findings suggest that explicit biomechanical phase information is a useful inductive prior for pose-based performance modeling in elite athletics throwing. Beyond score prediction, the dataset and protocol provide a basis for future work on fine-grained technique analysis, interpretable sports analytics, and robust modeling from broadcast video.