Research map / Video / Action recognition
Hierarchical Action Learning for Weakly-Supervised Action Segmentation
Junxian Huang, Ruichu Cai, Juntao Fang, Hao Zhu, Boyan Xu, Weilin Chen, Zijian Li, Shenghua Gao
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 reconstruction loss Lr alone (ablation) | HAL · Breakfast split 1 | IoU: 40.1 -> 38.4 | Hurt |
| Add reconstruction loss Lr alone (ablation) | HAL · Breakfast split 1 | MoF: 53.3 -> 54.3 | Improved |
| Add smoothness transition loss Ls alone (ablation) | HAL · Breakfast split 1 | MoF: 53.3 -> 54.6 | Improved |
| Add KL loss LKL alone (ablation) | HAL · Breakfast split 1 | IoU: 40.1 -> 42.0 | Improved |
| Add delta term alone (ablation) | HAL · Breakfast split 1 | MoF: 53.3 -> 53.9 | No clear effect |
| Combine Lr, Ls, LKL and delta term (ablation) | HAL · Breakfast split 1 | MoF: 53.3 -> 56.6 | Improved |
| Segment on high-level action variables instead of visual variables (HAL-V) (ablation) | HAL · Breakfast P04-cam01-P04-pancake | MoF: 36.2 -> 79.9 | Improved |
| Use HAL instead of ATBA | weakly-supervised action segmentation · Breakfast | MoF: 53.9 -> 56.3 | Improved |
| Use HAL instead of ATBA | weakly-supervised action segmentation · CrossTask | IoD: 44.6 -> 41.9 | Hurt |
| Use HAL instead of CtrlNS | weakly-supervised action segmentation · Hollywood Extended | MoF: 52.9 -> 51.0 | Hurt |
| Use HAL instead of ATBA | weakly-supervised action segmentation · GTEA | MoF: 42.2 -> 45.2 | Improved |
Closest papers by meaning
- Action Sets: Weakly Supervised Action Segmentation Without Ordering Constraints (CVPR 2018 · similarity 0.90)
- Weakly-Supervised Action Segmentation With Iterative Soft Boundary Assignment (CVPR 2018 · similarity 0.89)
- SCT: Set Constrained Temporal Transformer for Set Supervised Action Segmentation (CVPR 2020 · similarity 0.89)
- Weakly-Supervised Temporal Action Detection for Fine-Grained Videos with Hierarchical Atomic Actions (ECCV 2022 · similarity 0.89)
- Weakly-Supervised Action Localization by Hierarchically-Structured Latent Attention Modeling (ICCV 2023 · similarity 0.89)
- Efficient and Effective Weakly-Supervised Action Segmentation via Action-Transition-Aware Boundary Alignment (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.
