Automated Detection of Match Phases in Football from Spatio-Temporal Tracking Data Using Graph Neural Networks

Name: Automated Detection of Match Phases in Football from Spatio-Temporal Tracking Data Using Graph Neural Networks
Uploaded: 2026-10-14T16:07:37.830507+00:00
Duration: 903 s
Description: This work proposes a framework combining Graph Neural Networks (GNNs) with a sequential model to classify football match phases second-by-second across seven categories. By modeling player positions as a graph, the method outperforms all standard feature-based baselines. It shows that spatial relationships between players are crucial for accurately labeling tactical moments in a game.

This work proposes a framework combining Graph Neural Networks (GNNs) with a sequential model to classify football match phases second-by-second across seven categories. By modeling player positions as a graph, the method outperforms all standard feature-based baselines. It shows that spatial relationships between players are crucial for accurately labeling tactical moments in a game.

Open the full episode