Research map / Video / Action recognition
Improving Action Localization by Progressive Cross-Stream Cooperation
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 progressive cross-stream cooperation | action localization with I3D · UCF-101-24 frame level, IoU 0.5 | mAP (%): 75.5 -> 79.2 | Helped |
| use progressive cross-stream cooperation and temporal refinement | action localization with I3D · UCF-101-24 video level, IoU 0.5 | mAP (%): 53.2 -> 61.0 | Helped |
| use progressive cross-stream cooperation | action localization with I3D · J-HMDB frame level, IoU 0.5 | mAP (%): 77.9 -> 74.8 | Hurt |
| use progressive cross-stream cooperation | action localization with I3D · J-HMDB video level, IoU 0.5 | mAP (%): 78.5 -> 82.2 | Helped |
| add feature-level cooperation between streams (ablation) | PCSC · mini-UCF-101-24, Stage 0 | mAP (%): 72.7 -> 75.7 | Helped |
| add feature-level cooperation between streams (ablation) | PCSC · mini-UCF-101-24, Stage 1 | mAP (%): 73.3 -> 76.1 | Helped |
| add feature-level cooperation between streams (ablation) | PCSC · mini-UCF-101-24, Stage 2 | mAP (%): 73.5 -> 76.4 | Helped |
| add feature-level cooperation between streams (ablation) | PCSC · mini-UCF-101-24, Stage 3 | mAP (%): 73.9 -> 76.6 | Helped |
| add feature-level cooperation between streams (ablation) | PCSC · mini-UCF-101-24, Stage 4 | mAP (%): 73.9 -> 76.7 | Helped |
| add successive cross-stream cooperation stages (ablation) | PCSC · mini-UCF-101-24 frame level, Stage 0 to Stage 4 | mAP (%): 75.7 -> 76.7 | Helped |
| add class-specific actionness temporal refinement (ablation) | PCSC · mini-UCF-101-24 video level, IoU 0.5 | mAP (%): 55.8 -> 59.4 | Helped |
Closest papers by meaning
- STEP: Spatio-Temporal Progressive Learning for Video Action Detection (CVPR 2019 · similarity 0.89)
- Rethinking the Faster R-CNN Architecture for Temporal Action Localization (CVPR 2018 · similarity 0.89)
- One-Shot Action Localization by Learning Sequence Matching Network (CVPR 2018 · similarity 0.89)
- Learning To Refactor Action and Co-Occurrence Features for Temporal Action Localization (CVPR 2022 · similarity 0.88)
- Temporal Action Localization with Cross Layer Task Decoupling and Refinement (AAAI 2025 · similarity 0.87)
- End-to-End Spatio-Temporal Action Localisation with Video Transformers (CVPR 2024 · similarity 0.87)
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.