Research map / Video / Action recognition
Semi-Weakly-Supervised Learning of Complex Actions From Instructional Task Videos
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 |
|---|---|---|---|
| Use SRE loss and self-training instead of weakly supervised learning with self-training alone | MuCon · Breakfast, 1% weakly labeled | MoF: 22.3 -> 25.0 | Helped |
| Use SRE loss and self-training instead of weakly supervised learning with self-training alone | MuCon · CrossTask, 1% weakly labeled | MoF: 40.5 -> 48.1 | Helped |
| Use SRE loss and self-training instead of weakly supervised learning with self-training alone | CDFL · Breakfast, 1% weakly labeled | MoF: 18.6 -> 32.4 | Helped |
| Use SRE loss and self-training instead of weakly supervised learning with self-training alone | CDFL · CrossTask, 1% weakly labeled | MoF: 27.8 -> 21.8 | Hurt |
| Add self-training to SWSL (ablation) | MuCon · Breakfast, 1% weakly labeled | IoU: 21.8 -> 29.8 | Helped |
| Add self-training to SWSL (ablation) | CDFL · Breakfast, 1% weakly labeled | MoF: 13.4 -> 32.4 | Helped |
| Add adjacent transposition to the Soft Edit loss (ablation) | MuCon SWSL on Breakfast · Breakfast, 1% weakly labeled | MoF: 24.3 -> 25.0 | Helped |
| Add adjacent transposition to the Soft Edit loss (ablation) | MuCon SWSL on Breakfast · Breakfast, 1% weakly labeled | IoU: 29.2 -> 29.8 | Helped |
| Add adjacent transposition to the Soft Edit loss (ablation) | MuCon SWSL on Breakfast · Breakfast, 10% weakly labeled | IoU: 33.4 -> 33.4 | No clear effect |
| Add unlabeled videos using the SWSL loss (ablation) | MuCon · Breakfast, 1% weakly labeled | MoF: 11.0 -> 23.1 | Helped |
Closest papers by meaning
- Weakly-Supervised Online Action Segmentation in Multi-View Instructional Videos (CVPR 2022 · similarity 0.90)
- Large-Scale Weakly-Supervised Pre-Training for Video Action Recognition (CVPR 2019 · similarity 0.90)
- Action Sets: Weakly Supervised Action Segmentation Without Ordering Constraints (CVPR 2018 · similarity 0.90)
- Weakly Supervised Action Learning With RNN Based Fine-To-Coarse Modeling (CVPR 2017 · similarity 0.89)
- Efficient and Effective Weakly-Supervised Action Segmentation via Action-Transition-Aware Boundary Alignment (CVPR 2024 · similarity 0.89)
- Collaborative Weakly Supervised Video Correlation Learning for Procedure-Aware Instructional Video Analysis (AAAI 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.