Research map / Video / Action recognition
Segregated Temporal Assembly Recurrent Networks for Weakly Supervised Multiple Action Detection
Read the paperCheck 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 STAR instead of the strongest reported weakly supervised baseline | Weakly supervised action detection · THUMOS'14 | AP@IoU 0.1: 55.2 -> 68.8 | Helped |
| Use STAR instead of the strongest reported weakly supervised baseline | Weakly supervised action detection · ActivityNet1.3 validation | AP@IoU 0.5: 29.3 -> 31.1 | Helped |
| Add sparsity to the stem (ablation) | STAR stem · THUMOS'14 | Ave-mAP: 39.0 -> 43.8 | Helped |
| Add coverage to the stem (ablation) | STAR stem · THUMOS'14 | Ave-mAP: 39.0 -> 42.4 | Helped |
| Add RAM to the stem (ablation) | STAR stem · THUMOS'14 | Ave-mAP: 39.0 -> 43.8 | Helped |
| Add sparsity and coverage together (ablation) | STAR stem · THUMOS'14 | Ave-mAP: 43.8 -> 43.3 | Hurt |
| Add sparsity and RAM together (ablation) | STAR stem · THUMOS'14 | Ave-mAP: 39.0 -> 44.0 | Helped |
| Add coverage and RAM together (ablation) | STAR stem · THUMOS'14 | Ave-mAP: 39.0 -> 44.7 | Helped |
| Add sparsity, coverage, and RAM (ablation) | STAR stem · THUMOS'14 | Ave-mAP: 39.0 -> 47.0 | Helped |
| Fuse attention weights with ST-GradCAM rather than using attention alone (ablation) | STAR proposal localization · THUMOS'14 | Ave-mAP: 39.6 -> 47.0 | Helped |
| Fuse attention weights with ST-GradCAM rather than using ST-GradCAM alone (ablation) | STAR proposal localization · THUMOS'14 | Ave-mAP: 24.4 -> 47.0 | Helped |
Closest papers by meaning
- SCT: Set Constrained Temporal Transformer for Set Supervised Action Segmentation (CVPR 2020 · similarity 0.89)
- Weakly Supervised Action Learning With RNN Based Fine-To-Coarse Modeling (CVPR 2017 · similarity 0.88)
- Two-Stream Networks for Weakly-Supervised Temporal Action Localization With Semantic-Aware Mechanisms (CVPR 2023 · similarity 0.88)
- Weakly-Supervised Action Segmentation With Iterative Soft Boundary Assignment (CVPR 2018 · similarity 0.88)
- Action Unit Memory Network for Weakly Supervised Temporal Action Localization (CVPR 2021 · similarity 0.88)
- Temporal Convolutional Networks for Action Segmentation and Detection (CVPR 2017 · 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.