Research map / Video / Action recognition
A flexible model for training action localization with varying levels of supervision
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 temporal-point labels instead of temporal bounds (ablation) | Our method · DALY | mAP@0.2: 31.5 -> 26.7 | Hurt |
| Use temporal-point labels instead of temporal bounds (ablation) | Our method · UCF101-24 | mAP@0.2: 47.3 -> 45.5 | Hurt |
| Add a few fully supervised videos to video-level-labeled training videos | Our method · DALY | mAP@0.2: 7.6 -> 18.2 | Helped |
| Use temporal-point labels for the weakly labeled portion when mixing supervision | Our method · DALY | mAP@0.2: null | Helped |
| Add spatial bounding-box supervision to temporal bounds (ablation) | Our method · DALY | mAP@0.2: +1.0% | Helped |
| Add three sparse bounding boxes to temporal bounds (ablation) | Our method · UCF101-24 | mAP@0.2: 47.3 -> 74.5 | Helped |
| Add three sparse bounding boxes to temporal bounds (ablation) | Our method · UCF101-24 | mAP@0.5: 20.1 -> 43.2 | Helped |
| Use the cited method's tracks instead of generic person-detector tracks (ablation) | Our method · UCF101-24 video-level supervision | mAP@0.2: 43.9 -> 53.1 | Helped |
| Add one bounding box to temporal bounds (ablation) | Our method · UCF101-24 | mAP@0.2: 47.3 -> 70.6 | Helped |
Closest papers by meaning
- PivoTAL: Prior-Driven Supervision for Weakly-Supervised Temporal Action Localization (CVPR 2023 · similarity 0.92)
- Weakly Supervised Temporal Action Localization Through Learning Explicit Subspaces for Action and Context (AAAI 2021 · similarity 0.90)
- Weakly-supervised Temporal Action Localization by Uncertainty Modeling (AAAI 2021 · similarity 0.90)
- ASM-Loc: Action-Aware Segment Modeling for Weakly-Supervised Temporal Action Localization (CVPR 2022 · similarity 0.90)
- Weakly-Supervised Temporal Action Localization by Inferring Salient Snippet-Feature (AAAI 2024 · similarity 0.90)
- Similar Modality Enhancement and Action Consistency Learning for Weakly Supervised Temporal Action Localization (AAAI 2025 · similarity 0.89)
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.