Research map / Video / Video-language understanding
Learning Fine-grained View-Invariant Representations from Unpaired Ego-Exo Videos via Temporal Alignment
Zihui Xue, Kristen Grauman
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 |
|---|---|---|---|
| Add object-centric encoder to Base DTW (ablation) | DTW alignment with ResNet-50 · Break Eggs | F1 score: 58.53 -> 62.86 | Improved |
| Add object-centric encoder to Base DTW (ablation) | DTW alignment with ResNet-50 · Pour Milk | F1 score: 82.91 -> 84.04 | Improved |
| Add object-centric encoder to Base DTW (ablation) | DTW alignment with ResNet-50 · Pour Liquid | F1 score: 59.66 -> 63.28 | Improved |
| Add object-centric encoder to Base DTW (ablation) | DTW alignment with ResNet-50 · Tennis Forehand | F1 score: 79.56 -> 84.14 | Improved |
| Add contrastive regularization on top of object-centric encoder (ablation) | AE2 · Break Eggs | F1 score: 62.86 -> 66.23 | Improved |
| Add contrastive regularization on top of object-centric encoder (ablation) | AE2 · Pour Milk | F1 score: 84.04 -> 85.17 | Improved |
| Add contrastive regularization on top of object-centric encoder (ablation) | AE2 · Pour Liquid | F1 score: 63.28 -> 66.56 | Improved |
| Add contrastive regularization on top of object-centric encoder (ablation) | AE2 · Tennis Forehand | F1 score: 84.14 -> 85.87 | Improved |
| Add object-centric encoder to Base DTW (ablation) | DTW alignment with ResNet-50 · Pour Liquid | frame retrieval mAP@10: +10.83% | Improved |
| Add object-centric encoder to Base DTW (ablation) | DTW alignment with ResNet-50 · Tennis Forehand | frame retrieval mAP@10: +5.57% | Improved |
| Use random shuffle instead of temporal reverse for negative samples (ablation) | AE2 · Break Eggs | F1 score: -5.62% | Hurt |
| Use random shuffle instead of temporal reverse for negative samples (ablation) | AE2 · Break Eggs | frame retrieval mAP@10: 65.85 -> 61.66 | Hurt |
Closest papers by meaning
- Bootstrap Your Own Views: Masked Ego-Exo Modeling for Fine-grained View-invariant Video Representations (CVPR 2025 · similarity 0.92)
- Viewpoint Rosetta Stone: Unlocking Unpaired Ego-Exo Videos for View-invariant Representation Learning (CVPR 2025 · similarity 0.91)
- Modeling Fine-Grained Hand-Object Dynamics for Egocentric Video Representation Learning (ICLR 2025 · similarity 0.89)
- Action Scene Graphs for Long-Form Understanding of Egocentric Videos (CVPR 2024 · similarity 0.89)
- Ego-Exo4D: Understanding Skilled Human Activity from First- and Third-Person Perspectives (CVPR 2024 · similarity 0.88)
- Retrieval-Augmented Egocentric Video Captioning (CVPR 2024 · similarity 0.88)
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.
