Research map / Video / Action recognition
Localizing the Common Action Among a Few Videos
Pengwan Yang, Vincent Tao Hu, Pascal Mettes, Cees G. M. Snoek
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 progressive alignment module (PAM) (ablation) | Baseline without modules · ActivityNet common instance (one-shot) | mAP: 42.4 -> 49.7 | Improved |
| add progressive alignment module (PAM) (ablation) | Baseline without modules · Thumos common instance (five-shot) | mAP: 38.4 -> 44.5 | Improved |
| add pairwise matching module (PMM) on top of PAM (ablation) | PAM · ActivityNet common instance (one-shot) | mAP: 49.7 -> 51.3 | Improved |
| add mutual enhancement module (MEM) before PAM (ablation) | PAM · ActivityNet common instance (five-shot) | mAP: 52.0 -> 55.3 | Improved |
| add PMM on top of MEM + PAM (ablation) | MEM + PAM · Thumos common instance (five-shot) | mAP: 49.6 -> 51.9 | Improved |
| add all three modules (MEM, PAM, PMM) (ablation) | Baseline · ActivityNet common instance (one-shot) | mAP: 42.4 -> 53.1 | Improved |
| increase the number of support videos from one to five | Full model · ActivityNet common instance (shot count) | mAP: 53.1 -> 56.5 | Improved |
| increase support videos from five to six | Full model · ActivityNet common instance (shot count) | mAP: 56.5 -> 56.8 | No clear effect |
| replace one of five support videos with a wrong-action video | Full model · ActivityNet common instance (five-shot) | mAP: 56.5 -> 53.5 | Hurt |
| replace two of five support videos with noisy videos of the same class | Full model · ActivityNet common instance (five-shot) | mAP: 56.5 -> 50.6 | Hurt |
| increase PAM depth from 1 to 3 (ablation) | PAM · ActivityNet common instance (five-shot) | mAP: 53.9 -> 56.5 | Improved |
| increase PAM depth from 3 to 5 (ablation) | PAM · ActivityNet common instance (five-shot) | mAP: 56.5 -> 56.4 | No clear effect |
Closest papers by meaning
- Few-Shot Transformation of Common Actions Into Time and Space (CVPR 2021 · similarity 0.95)
- Temporal Action Localization in Untrimmed Videos via Multi-Stage CNNs (CVPR 2016 · similarity 0.92)
- Graph Convolutional Module for Temporal Action Localization in Videos (TPAMI 2021 · similarity 0.91)
- Few-Shot Common Action Localization via Cross-Attentional Fusion of Context and Temporal Dynamics (ICCV 2023 · similarity 0.91)
- CDC: Convolutional-De-Convolutional Networks for Precise Temporal Action Localization in Untrimmed Videos (CVPR 2017 · similarity 0.91)
- Weakly-Supervised Temporal Action Localization by Inferring Salient Snippet-Feature (AAAI 2024 · 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.
