Research map / Video / Action recognition
Learning by Aligning Videos in Time
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 Contrastive-IDM temporal regularization to Soft-DTW (ablation) | LAV · Pouring | phase progression: 0.2770 -> 0.8054 | Helped |
| Add Contrastive-IDM temporal regularization to Soft-DTW (ablation) | LAV · Pouring | Kendall's Tau: 0.2144 -> 0.8561 | Helped |
| Add IDM regularization to Soft-DTW instead of Contrastive-IDM (ablation) | LAV · Pouring | phase progression: 0.8054 vs. 0.6551 | Hurt |
| Add SFA regularization to Soft-DTW instead of Contrastive-IDM (ablation) | LAV · Pouring | Kendall's Tau: 0.8561 vs. 0.8069 | Hurt |
| Use Contrastive-IDM rather than IDM alone (ablation) | temporal regularization · Pouring | Kendall's Tau: 0.5835 -> 0.8318 | Helped |
| Use Contrastive-IDM rather than SFA alone (ablation) | temporal regularization · Pouring | Kendall's Tau: 0.8093 -> 0.8318 | Helped |
| Combine LAV with TCC | phase classification · Pouring | phase classification accuracy (1.0% labels): 92.84 -> 93.07 | Helped |
| Combine LAV with TCC | phase classification · IKEA ASM | phase classification accuracy (1.0% labels): 30.43 -> 30.51 | No clear effect |
| Combine LAV with TCC | phase classification · Penn Action | phase classification accuracy (1.0% labels): 84.25 -> 84.12 | Hurt |
| Use LAV instead of TCC | video representation learning · Pouring | Average Precision at 5: 87.16 -> 89.13 | Helped |
| Use LAV instead of TCN | video representation learning · Penn Action | Average Precision at 5: 77.84 -> 79.13 | Helped |
Closest papers by meaning
- Temporal Cycle-Consistency Learning (CVPR 2019 · similarity 0.91)
- Learning To Align Sequential Actions in the Wild (CVPR 2022 · similarity 0.91)
- Temporal Alignment Networks for Long-Term Video (CVPR 2022 · similarity 0.91)
- No More Shortcuts: Realizing the Potential of Temporal Self-Supervision (AAAI 2024 · similarity 0.89)
- DejaVid: Encoder-Agnostic Learned Temporal Matching for Video Classification (CVPR 2025 · similarity 0.89)
- Align Before Adapt: Leveraging Entity-to-Region Alignments for Generalizable Video Action Recognition (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.