Research map / Video / Action recognition
EgoExo-Fitness: Towards Egocentric and Exocentric Full-Body Action Understanding
Yuan-Ming Li, Wei-Jin Huang, An-Lan Wang, Ling-An Zeng, Jing-Ke Meng, Wei-Shi Zheng
Read the paperCheck your idea against 137,217 papers
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 |
|---|---|---|---|
| train on mixed ego and exo data instead of single view | TimeSformer (K600) · Action classification, Ego test | Top-1 accuracy: 0.7932 -> 0.7842 | Hurt |
| train on mixed ego and exo data instead of single view | I3D (K400) · Action classification, Exo test | Top-1 accuracy: 0.9194 -> 0.8963 | Hurt |
| train ego-only or exo-only model and test on the other view | I3D (K400) · Action classification | Top-1 accuracy: 0.9194 -> 0.0927 (Exo-trained, Ego test) | Hurt |
| view-matched pretraining (Ego-Exo4D vs Kinetics) for egocentric test | TimeSformer · Action classification, Ego test | Top-1 accuracy: 0.7932 -> 0.8000 | Improved |
| add synchronized video alignment loss L_Align (ablation) | GEVFormer (ego+exo) · Guidance-based execution verification, Ego | F1-score: 0.5248 -> 0.5425 | Improved |
| add L_Align (ablation) | GEVFormer (ego+exo) · Guidance-based execution verification, Exo | F1-score: 0.5282 -> 0.5452 | Mixed |
| train on ego+exo jointly without alignment instead of separate models (ablation) | GEVFormer · Guidance-based execution verification | F1-score: Ego +0.0087, Exo -0.0192 | Mixed |
| add Temporal Context Modeling module to CLIP-GEV (ablation) | CLIP-GEV · Guidance-based execution verification | average F1-score: 0.4881 -> 0.5138 | Improved |
| add Cross-Modal Verifier decoder (ablation) | CLIP-GEV + TCM · Guidance-based execution verification | average F1-score: 0.5138 -> 0.5265 | Improved |
| jointly train localization on ego and exo data | TadTR · Action localization, Exo test | mAP: Ego 33.66 -> 35.63; Exo 40.92 -> 39.77 | Mixed |
Closest papers by meaning
- Ego-Exo4D: Understanding Skilled Human Activity from First- and Third-Person Perspectives (CVPR 2024 · similarity 0.90)
- Action Scene Graphs for Long-Form Understanding of Egocentric Videos (CVPR 2024 · similarity 0.88)
- EgoSim: An Egocentric Multi-view Simulator and Real Dataset for Body-worn Cameras during Motion and Activity (NeurIPS 2024 · similarity 0.88)
- EgoExoLearn: A Dataset for Bridging Asynchronous Ego- and Exo-centric View of Procedural Activities in Real World (CVPR 2024 · similarity 0.87)
- EgoVid-5M: A Large-Scale Video-Action Dataset for Egocentric Videos Generation (NeurIPS 2025 · similarity 0.87)
- Retrieval-Augmented Egocentric Video Captioning (CVPR 2024 · similarity 0.87)
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.
