Entity-Aware Sequence Transduction for Player-Centric Ball Action Spotting

2026/08/03 by Ruifeng Wang, Di Yang, Jiangtao Wang
Computer Science · #cs.CV #cs.AI

paper · pdf

arxiv created 2026/08/03 · arxiv updated 2026/08/04

Abstract

Player-centric ball action spotting requires temporally precise event detection together with actor attribution in crowded, partially observed multi-agent sports videos. Existing Denoising Sequence Transduction (DST) baselines treat the player-role dimension as part of a flattened frame-level representation, which weakens the inductive bias for modeling player-specific temporal evolution and inter-player interactions. To address this limitation, we propose Multi-Entity Denoising Sequence Transduction (ME-DST). ME-DST keeps the role-slot dimension throughout encoding. It uses temporal attention to model the history of each role slot, and spatial attention to exchange information across role slots at each frame. This factorized design gives the model a direct structure for separating within-player evolution from inter-player context. We also add learnable role embeddings, tracking-derived tactical features, and fused visual predictions from X3D-L and Swin3D-S. Experiments on the FOOTPASS dataset show that ME-DST reaches a Micro F1 of 0.778. This improves the strongest official TAAD+DST baseline by 10.3 percentage points. Controlled ablations show that preserving the entity axis and encoding role identity are central to this gain. These results suggest that explicit entity modeling is an effective inductive bias for player-centric sports event understanding.

Citations

• FOOTPASS: A Multi-Modal Multi-Agent Tactical Context Dataset for Play-by-Play Action Spotting in Soccer Broadcast Videos
• SoccerNet 2025 Challenges Results
• Beyond Pixels: Leveraging the Language of Soccer to Improve Spatio-Temporal Action Detection in Broadcast Videos
• Action Anticipation from SoccerNet Football Video Broadcasts
• Precise Event Spotting in Sports Videos: Solving Long-Range Dependency and Class Imbalance
• Game State and Spatio-temporal Action Detection in Soccer using Graph Neural Networks and 3D Convolutional Networks
• SoccerNet 2024 Challenges Results
• T-DEED: Temporal-Discriminability Enhancer Encoder-Decoder for Precise Event Spotting in Sports Videos
• ASTRA: An Action Spotting TRAnsformer for Soccer Videos
• COMEDIAN: Self-Supervised Learning and Knowledge Distillation for Action Spotting using Transformers
• A Boosted Model Ensembling Approach to Ball Action Spotting in Videos: The Runner-Up Solution to CVPR'23 SoccerNet Challenge
• Spatio-Temporal Action Detection Under Large Motion
• Spotting Temporally Precise, Fine-Grained Events in Video
• Temporally Precise Action Spotting in Soccer Videos Using Dense Detection Anchors
• GroupFormer: Group Activity Recognition with Clustered Spatial-Temporal Transformer
• Video Swin Transformer
• Video Swin Transformer
• Temporally-Aware Feature Pooling for Action Spotting in Soccer\n Broadcasts
• EfficientNetV2: Smaller Models and Faster Training
• AgentFormer: Agent-Aware Transformers for Socio-Temporal Multi-Agent Forecasting
• SoccerNet-v2: A Dataset and Benchmarks for Holistic Understanding of\n Broadcast Soccer Videos
• X3D: Expanding Architectures for Efficient Video Recognition
• Actor-Transformers for Group Activity Recognition
• A Context-Aware Loss Function for Action Spotting in Soccer Videos
• Attention Is All You Need
• Training Deep Nets with Sublinear Memory Cost
• Rethinking the Inception Architecture for Computer Vision
• NetVLAD: CNN architecture for weakly supervised place recognition
• Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation