Technology Category
Computer Vision: Motion & TrackingIntelligent Robots: Multimodal Perception & Sensor FusionMachine Learning: Multi-instance/Multi-view Learning
SoccerTrack v2: A Full-Pitch Multi-View Soccer Dataset for Game State Reconstruction
🤖 AI Summary
Existing football analysis datasets are largely constrained to broadcast views or localized scenes, limiting joint research on multi-object tracking (MOT), game state reconstruction (GSR), and ball action segmentation (BAS). To address this, we introduce the first GSR-oriented, full-field, multi-view, 4K-resolution football dataset, comprising 10 complete university-level matches. Leveraging BePro panoramic cameras, we capture six synchronized video streams per match and employ a hybrid manual–semi-automatic annotation pipeline. The dataset provides frame-level 2D on-field coordinates, player IDs and roles, and 12 fine-grained ball possession action labels. By overcoming traditional viewpoint and scene limitations, it significantly enhances player visibility and state reconstruction fidelity. This unified benchmark enables rigorous evaluation and co-development of MOT, GSR, and BAS methods, thereby advancing computer vision applications in tactical football analysis.
Technology Category
Application Category
📝 Abstract
SoccerTrack v2 is a new public dataset for advancing multi-object tracking (MOT), game state reconstruction (GSR), and ball action spotting (BAS) in soccer analytics. Unlike prior datasets that use broadcast views or limited scenarios, SoccerTrack v2 provides 10 full-length, panoramic 4K recordings of university-level matches, captured with BePro cameras for complete player visibility. Each video is annotated with GSR labels (2D pitch coordinates, jersey-based player IDs, roles, teams) and BAS labels for 12 action classes (e.g., Pass, Drive, Shot). This technical report outlines the datasets structure, collection pipeline, and annotation process. SoccerTrack v2 is designed to advance research in computer vision and soccer analytics, enabling new benchmarks and practical applications in tactical analysis and automated tools.
Problem
Research questions and friction points this paper is trying to address.
Advance multi-object tracking in soccer analytics
Enable game state reconstruction with full-pitch data
Improve ball action spotting with detailed annotations
Innovation
Methods, ideas, or system contributions that make the work stand out.
Panoramic 4K recordings for full-pitch visibility
GSR and BAS labels for detailed analytics
BePro cameras ensure complete player tracking
🔎 Similar Papers
No similar papers found.
