Research map / Video / Action recognition
Fast Temporal Activity Proposals for Efficient Detection of Human Actions in Untrimmed 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 |
|---|---|---|---|
| add activity-label supervision to dictionary learning (ablation) | Temporal activity proposals · MSR-II | recall: Class-Induced: 80%; Class-Independent: 60.3 | Helped |
| add activity-label supervision to dictionary learning (ablation) | Temporal activity proposals · Thumos14 | mAP (tIoU 0.5, 1000 proposals): 6.7 -> 13.5 | Helped |
| use class-induced proposals instead of uniform sampling | Action detection · Thumos14 | mAP (tIoU 0.5, 1000 proposals): 3.1 -> 13.5 | Helped |
| use class-induced proposals instead of BPC proposals | Action detection · Thumos14 | mAP (tIoU 0.5, 1000 proposals): 9.9 -> 13.5 | Helped |
| use class-induced proposals instead of APT proposals | Action detection · Thumos14 | mAP (tIoU 0.5, 1000 proposals): 9.1 -> 13.5 | Helped |
| use class-induced proposals instead of class-independent proposals (ablation) | Action detection · MSR-II | mAP (tIoU 0.5, 100 proposals): 34.1 -> 60.3 | Helped |
| use class-induced proposals instead of class-independent proposals (ablation) | Action detection · Thumos14 | mAP (tIoU 0.8, 1000 proposals): 3.3 -> 7.5 | Helped |
| use class-induced proposals instead of APT | Temporal proposal generation · Thumos14 | FPS: 0.68 -> 10.2 | Helped |
| use class-induced proposals instead of Oneata et al. | Action detection · Thumos14 detection challenge | mAP (tIoU 0.5): 14.3% -> 13.5% | Hurt |
| use class-induced proposals instead of FAP | Temporal proposals · MSR-II | recall at high tIoU: At tIoU 0.5, FAP recall is 0.6; exact class-induced value not stated in text. | Helped |
Closest papers by meaning
- Action Recognition From Single Timestamp Supervision in Untrimmed Videos (CVPR 2019 · similarity 0.90)
- PointTAD: Multi-Label Temporal Action Detection with Learnable Query Points (NeurIPS 2022 · similarity 0.89)
- Unsupervised Action Segmentation by Joint Representation Learning and Online Clustering (CVPR 2022 · similarity 0.88)
- Unsupervised Learning of Action Classes With Continuous Temporal Embedding (CVPR 2019 · similarity 0.88)
- Three Birds with One Stone: Multi-Task Temporal Action Detection via Recycling Temporal Annotations (CVPR 2021 · similarity 0.88)
- Action-Agnostic Point-Level Supervision for Temporal Action Detection (AAAI 2025 · 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.