Research map / Video / Action recognition
Activity Grammars for Temporal Action Segmentation
Dayoung Gong, Joonseok Lee, Deunsol Jung, Suha Kwak, Minsu Cho
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 KARI grammar refinement | ASFormer · 50Salads | edit score: 76.5 -> 79.9 | Improved |
| add KARI grammar refinement | ASFormer · 50Salads | accuracy: 86.1 -> 85.3 | Hurt |
| add KARI grammar refinement | ASFormer · Breakfast | F1@50: 59.4 -> 60.8 | Improved |
| add KARI grammar refinement | MS-TCN · 50Salads | edit score: 62.4 -> 66.7 | Improved |
| add KARI grammar refinement | MS-TCN · Breakfast | accuracy: 69.4 -> 68.8 | No clear effect |
| use ADIOS-OR grammar refinement instead of none | ASFormer · 50Salads | edit score: 76.5 -> 61.1 | Hurt |
| use KARI instead of ADIOS-OR grammar | ASFormer · 50Salads | edit score: 61.1 -> 79.9 | Improved |
| add recursive rules to KARI (ablation) | KARI · 50Salads | edit score: 69.2 -> 79.9 | Improved |
| add key actions to KARI (ablation) | KARI · 50Salads | edit score: 62.6 -> 69.2 | Improved |
| use BEP parser instead of GEP (ablation) | KARI · 50Salads | edit score: 72.1 -> 79.9 | Improved |
| vary queue size N_queue 10 to 30 for BEP (ablation) | BEP · 50Salads | edit score: 78.3 -> 78.9 | No clear effect |
| number of key actions N_key from 3 to 6 (ablation) | KARI · 50Salads | edit score: 79.9 -> 68.4 | Hurt |
Closest papers by meaning
- Action Sequence Augmentation for Action Anticipation (ICLR 2025 · similarity 0.89)
- OnlineTAS: An Online Baseline for Temporal Action Segmentation (NeurIPS 2024 · similarity 0.88)
- Temporal Action Segmentation From Timestamp Supervision (CVPR 2021 · similarity 0.88)
- Unsupervised Action Segmentation via Fast Learning of Semantically Consistent Actoms (AAAI 2024 · similarity 0.87)
- FACT: Frame-Action Cross-Attention Temporal Modeling for Efficient Action Segmentation (CVPR 2024 · similarity 0.87)
- ActFusion: a Unified Diffusion Model for Action Segmentation and Anticipation (NeurIPS 2024 · similarity 0.86)
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.
