Research map / Video / Action recognition
Progressive Cross-Stream Cooperation in Spatial and Temporal Domain for Action Localization
Rui Su, Dong Xu, Luping Zhou, Wanli Ouyang
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 cross-stream cooperation to Faster R-CNN + FPN | two-stream detector · UCF-101-24 frame-level | frame mAP (%): 75.5 -> 79.2 | Improved |
| add cross-stream cooperation to Faster R-CNN + FPN | two-stream detector · J-HMDB-21 video-level IoU 0.5 | video mAP (%): 78.5 -> 82.2 | Improved |
| add feature-level cooperation on top of proposal cooperation (stage 4) (ablation) | PCSC · UCF-101-24 | frame mAP (%): 76.7 -> 79.2 | Improved |
| add progressive stages (stage 0 to stage 4) (ablation) | PCSC · UCF-101-24 | frame mAP (%): 78.1 -> 79.2 | Improved |
| add progressive stages (stage 0 to stage 4) without feature cooperation (ablation) | PCSC w/o feature cooperation · UCF-101-24 | frame mAP (%): 75.5 -> 76.7 | Improved |
| add actionness detector temporal refinement to PCSC (ablation) | PCSC · UCF-101-24 video-level IoU 0.5 | video mAP (%): 56.4 -> 61.0 | Improved |
| use multi-branch instead of single-branch action segment detector (ablation) | PCSC + ASD · UCF-101-24 video-level IoU 0.5 | video mAP (%): 58.4 -> 60.7 | Improved |
| combine actionness detector with action segment detector (ablation) | PCSC + ASD · UCF-101-24 video-level IoU 0.5 | video mAP (%): 60.7 -> 61.9 | Improved |
| add Temporal PCSC (two stages) (ablation) | PCSC + ASD + AD · UCF-101-24 video-level IoU 0.5 | video mAP (%): 61.9 -> 63.1 | Improved |
| add second T-PCSC stage (ablation) | T-PCSC · UCF-101-24 video-level IoU 0.5 | video mAP (%): 62.8 -> 63.1 | Improved |
| add progressive stages to PCSC (ablation) | PCSC · UCF-101-24 | large overlap ratio (%): 61.1 -> 95.4 | Mixed |
| add progressive stages to PCSC (ablation) | PCSC · UCF-101-24 | detection speed (FPS): 3.1 -> 2.4 | Hurt |
Closest papers by meaning
- Improving Action Localization by Progressive Cross-Stream Cooperation (CVPR 2019 · similarity 0.98)
- Graph Convolutional Module for Temporal Action Localization in Videos (TPAMI 2021 · similarity 0.90)
- Action Tubelet Detector for Spatio-Temporal Action Localization (ICCV 2017 · similarity 0.89)
- STEP: Spatio-Temporal Progressive Learning for Video Action Detection (CVPR 2019 · similarity 0.89)
- Pointly-Supervised Action Localization (IJCV 2018 · similarity 0.88)
- 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.
