Research map / Video / Video segmentation and tracking
Few-Shot Precise Event Spotting via Unified Multi-Entity Graph and Distillation
Zhaoyu Liu, Kan Jiang, Murong Ma, Zhe Hou, Yun Lin, Jin Song Dong
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add ball keypoint to the pose-only graph (ablation) | UMEG-Net · ShuttleSet (100-clip) | Edit score: 61.5 -> 62.5 | Improved |
| add court corners to the pose-only graph (ablation) | UMEG-Net · ShuttleSet (100-clip) | F1evt: 47.4 -> 46.7 | No clear effect |
| add ball and court to pose-only graph (ablation) | UMEG-Net · F3Set-Tennis (100-clip) | F1evt: 5.6 -> 9.4 | Improved |
| add ball and court to pose-only graph (ablation) | UMEG-Net · F3Set-Tennis (100-clip) | Edit score: 23.9 -> 31.7 | Improved |
| add ball keypoint to pose graph (ablation) | UMEG-Net · SoccerNet-BAS (100-clip) | F1evt: 20.7 -> 27.0 | Improved |
| temporal shift scales {1,2,4} instead of {1} (ablation) | UMEG-Net · FineGym-BB (100-clip) | F1evt: 39.0 -> 49.2 | Improved |
| temporal shift scales {1,2,4} instead of {1,2} (ablation) | UMEG-Net · F3Set-Tennis (100-clip) | F1evt: 9.6 -> 9.4 | Hurt |
| temporal shift scales {1,2,4} instead of {1,2} (ablation) | UMEG-Net · Figure Skating (100-clip) | F1evt: 32.1 -> 39.2 | Improved |
| multimodal distillation to RGB student | UMEG-Net · ShuttleSet (100-clip) | F1evt: 49.2 -> 59.1 | Improved |
| distillation on SoccerNet-BAS | UMEG-Net · SoccerNet-BAS (100-clip) | F1evt: 27.0 -> 27.1 | No clear effect |
| replace contrastive self-supervised pretraining with UMEG-Net distillation (ablation) | RGB student · F3Set-Tennis (100-clip) | F1evt: 3.0 -> 12.5 | Improved |
| contrastive self-supervision vs distillation on Shuttleset (ablation) | RGB student · ShuttleSet (100-clip) | F1evt: 50.2 -> 59.1 | Improved |
Closest papers by meaning
- Precise Event Spotting in Sports Videos: Solving Long-Range Dependency and Class Imbalance (CVPR 2025 · similarity 0.90)
- "Spotting Temporally Precise, Fine-Grained Events in Video" (ECCV 2022 · similarity 0.88)
- SportsCap: Monocular 3D Human Motion Capture and Fine-Grained Understanding in Challenging Sports Videos (IJCV 2021 · similarity 0.86)
- ViSTec: Video Modeling for Sports Technique Recognition and Tactical Analysis (AAAI 2024 · similarity 0.84)
- FineSports: A Multi-person Hierarchical Sports Video Dataset for Fine-grained Action Understanding (CVPR 2024 · similarity 0.84)
- A Dynamic GCN with Cross-Representation Distillation for Event-Based Learning (AAAI 2024 · similarity 0.83)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
