Research map / Video / Action recognition
AutoLoc: Weakly-supervised Temporal Action Localization in Untrimmed Videos
Zheng Shou, Hang Gao, Lei Zhang, Kazuyuki Miyazawa, Shih-Fu Chang
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 the outer-inner contrast to the loss rather than inner activation only (ablation) | AutoLoc · ActivityNet v1.2 val | mAP@IoU 0.5: 4.6 -> 27.3 | Improved |
| add the outer-inner contrast to the loss rather than inner activation only (ablation) | AutoLoc · ActivityNet v1.2 val | average mAP: 1.6 -> 16.0 | Improved |
| train the boundary predictor on training videos instead of direct optimization on each test video (ablation) | AutoLoc · ActivityNet v1.2 val | mAP@IoU 0.5: 21.8 -> 27.3 | Improved |
| train the boundary predictor on training videos instead of direct optimization on each test video (ablation) | AutoLoc · ActivityNet v1.2 val | average mAP: 12.6 -> 16.0 | Improved |
| use a learned boundary predictor instead of OIC Selection over all enumerated segments (ablation) | AutoLoc · ActivityNet v1.2 val | mAP@IoU 0.5: 15.8 -> 27.3 | Improved |
| use OIC-loss-based segment selection instead of CAS thresholding (ablation) | UntrimmedNet CAS · ActivityNet v1.2 val | mAP@IoU 0.5: 7.4 -> 15.8 | Improved |
| replace CAS thresholding with direct boundary prediction trained with OIC loss | UntrimmedNet · THUMOS'14 test | mAP@IoU 0.5: 13.7 -> 21.2 | Improved |
| replace CAS thresholding with direct boundary prediction trained with OIC loss | UntrimmedNet · ActivityNet v1.2 val | mAP@IoU 0.5: 7.4 -> 27.3 | Improved |
| weak-supervised AutoLoc versus full-supervised S-CNN | AutoLoc · THUMOS'14 test | mAP@IoU 0.5: 19.0 -> 21.2 | Improved |
Closest papers by meaning
- Weakly Supervised Temporal Action Localization Through Learning Explicit Subspaces for Action and Context (AAAI 2021 · similarity 0.92)
- Learning Temporal Co-Attention Models for Unsupervised Video Action Localization (CVPR 2020 · similarity 0.92)
- PivoTAL: Prior-Driven Supervision for Weakly-Supervised Temporal Action Localization (CVPR 2023 · similarity 0.92)
- Realigning Confidence with Temporal Saliency Information for Point-Level Weakly-Supervised Temporal Action Localization (CVPR 2024 · similarity 0.91)
- Bridge the Gap: From Weak to Full Supervision for Temporal Action Localization with PseudoFormer (CVPR 2025 · similarity 0.91)
- Weakly-Supervised Temporal Action Localization by Inferring Salient Snippet-Feature (AAAI 2024 · similarity 0.91)
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.
