Research map / Video / Action recognition
Exploring Denoised Cross-Video Contrast for Weakly-Supervised Temporal Action Localization
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 DCC to the baseline (ablation) | BaSNet baseline · THUMOS14 test | mAP@IoU 0.1:0.7 (Avg): 37.7 -> 44.0 | Helped |
| add pseudo-label denoising (ablation) | DCC without denoising · THUMOS14 test | mAP@IoU 0.1:0.7 (Avg): 42.3 -> 44.0 | Helped |
| vary K-means cluster count from 5 to 3 (ablation) | DCC · THUMOS14 test | mAP@IoU 0.1:0.7 (Avg): 44.0 -> 43.3 | Hurt |
| vary K-means cluster count from 5 to 10, 15, 50, or 100 (ablation) | DCC · THUMOS14 test | mAP@IoU 0.1:0.7 (Avg): 44.0 -> 43.8; 43.9; 43.5; 43.2 | Hurt |
| add inter-class contrast to action-background contrast (ablation) | video-level contrast model · THUMOS14 test | mAP@IoU 0.1:0.7 (Avg): 38.8 -> 39.6 | Helped |
| use region-level rather than video-level contrast features (ablation) | DCC with action-background and intra/inter-class contrast · THUMOS14 test | mAP@IoU 0.1:0.7 (Avg): 39.6 -> 42.3 | Helped |
| vary region number M from 5 to 1 or 3 (ablation) | DCC · THUMOS14 test | mAP@IoU 0.1:0.7 (Avg): 42.3 -> 39.6; 41.2 | Hurt |
| increase region number M from 5 to 10 (ablation) | DCC · THUMOS14 test | mAP@IoU 0.1:0.7 (Avg): 42.3 -> 41.7 | Hurt |
| add inter-video contrast without a memory bank (ablation) | intra-video contrast model · THUMOS14 test | mAP@IoU 0.1:0.7 (Avg): 38.5 -> 39.7 | Helped |
| add an entire-dataset region-level memory bank for inter-video contrast (ablation) | mini-batch inter-video contrast model · THUMOS14 test | mAP@IoU 0.1:0.7 (Avg): 39.7 -> 42.3 | Helped |
| add DCC (ablation) | STPN · THUMOS14 test | mAP@IoU 0.1:0.7 (Avg): 33.5 -> 40.5 | Helped |
| add DCC (ablation) | BaSNet · THUMOS14 test | mAP@IoU 0.1:0.7 (Avg): 37.7 -> 44.0 | Helped |
Closest papers by meaning
- Fine-Grained Temporal Contrastive Learning for Weakly-Supervised Temporal Action Localization (CVPR 2022 · similarity 0.93)
- CoLA: Weakly-Supervised Temporal Action Localization With Snippet Contrastive Learning (CVPR 2021 · similarity 0.92)
- Weakly-Supervised Temporal Action Localization by Inferring Salient Snippet-Feature (AAAI 2024 · similarity 0.92)
- PivoTAL: Prior-Driven Supervision for Weakly-Supervised Temporal Action Localization (CVPR 2023 · similarity 0.90)
- Cross-Attentional Audio-Visual Fusion for Weakly-Supervised Action Localization (ICLR 2021 · similarity 0.90)
- Bridge the Gap: From Weak to Full Supervision for Temporal Action Localization with PseudoFormer (CVPR 2025 · 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.