Research map / Video / Action recognition
Temporal Cycle-Consistency Learning
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 |
|---|---|---|---|
| Train embeddings by aligning frames across multiple videos with TCC | VGG-M from scratch · Penn Action, 10% labels | phase classification accuracy (%): 69.65 (TCN) -> 74.68 | Helped |
| Train embeddings by aligning frames across multiple videos with TCC | VGG-M from scratch · Pouring, 10% labels | phase classification accuracy (%): 76.03 (TCN) -> 86.82 | Helped |
| Combine TCC with TCN loss | ImageNet-pretrained ResNet-50 · Penn Action, 10% labels | phase classification accuracy (%): 81.99 (TCN) -> 84.27 | Helped |
| Combine TCC with TCN loss | ImageNet-pretrained ResNet-50 · Pouring, 10% labels | phase classification accuracy (%): 89.23 (TCC) -> 89.17 | No clear effect |
| Combine TCC with TCN loss | ImageNet-pretrained ResNet-50 · Penn Action, fine-tuning | Kendall's Tau: 0.7353 -> 0.7672 | Helped |
| Combine TCC with TCN loss | ImageNet-pretrained ResNet-50 · Pouring, fine-tuning | Kendall's Tau: 0.8516 -> 0.8779 | Helped |
| Add variance modeling to cycle-back regression (ablation) | TCC · Pouring | Kendall's Tau: 0.6093 -> 0.8516 | Helped |
| Use cycle-back regression instead of cycle-back classification (ablation) | TCC · Pouring | phase progression: 0.6636 -> 0.8030 | Helped |
Closest papers by meaning
- Representation Learning via Global Temporal Alignment and Cycle-Consistency (CVPR 2021 · similarity 0.92)
- Learning by Aligning Videos in Time (CVPR 2021 · similarity 0.91)
- Learning to Predict Activity Progress by Self-Supervised Video Alignment (CVPR 2024 · similarity 0.91)
- Learning To Align Sequential Actions in the Wild (CVPR 2022 · similarity 0.89)
- Temporal Alignment Networks for Long-Term Video (CVPR 2022 · similarity 0.88)
- TimeBalance: Temporally-Invariant and Temporally-Distinctive Video Representations for Semi-Supervised Action Recognition (CVPR 2023 · 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.