Research map / Video / Action recognition
Class Semantics-Based Attention for Action Detection
Deepak Sridhar, Niamul Quader, Srikanth Muralidharan, Yaoxin Li, Peng Dai, Juwei Lu
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 CSA attention | BMN · THUMOS-14 | avg mAP: 41.3 -> 47.5 | Improved |
| add CSA attention | GTAD · THUMOS-14 | avg mAP: 38.8 -> 42.6 | Improved |
| add CSA attention | BMN · ActivityNet v1.3 | avg mAP: 34.83 -> 35.43 | Improved |
| add CSA attention | BMN · ActivityNet v1.3 (high tIoU) | mAP@tIoU0.95: 6.53 -> 5.18 | Hurt |
| add CSA attention | DBG · ActivityNet v1.3 proposals | AUC: 68.19 -> 68.89 | Improved |
| use CSA instead of SE attention at encoder output (ablation) | BMN · THUMOS-14 | avg mAP: 44.2 -> 47.5 | Improved |
| use CSA instead of Transformer attention (ablation) | BMN · THUMOS-14 | avg mAP: 39.1 -> 47.5 | Improved |
| add 1D convolution in attention instead of feed-forward (ablation) | CSA · THUMOS-14 | avg mAP: 44.1 -> 46.7 | Improved |
| add temporal attention to channel attention (ablation) | CSA · THUMOS-14 | avg mAP: 45.1 -> 46.7 | Improved |
| add channel attention to temporal attention (ablation) | CSA · THUMOS-14 | avg mAP: 45.7 -> 46.7 | Improved |
| increase conv kernel size from 3 to 5 (ablation) | CSA · THUMOS-14 | avg mAP: 46.7 -> 39.8 | Hurt |
| increase conv blocks from 1 to 2 (ablation) | CSA · THUMOS-14 | avg mAP: 46.7 -> 47.5 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- GTAD — G-TAD: Sub-Graph Localization for Temporal Action Detection (CVPR 2020) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Graph Convolutional Module for Temporal Action Localization in Videos (TPAMI 2021 · similarity 0.92)
- One-Shot Action Localization by Learning Sequence Matching Network (CVPR 2018 · similarity 0.91)
- A Hybrid Attention Mechanism for Weakly-Supervised Temporal Action Localization (AAAI 2021 · similarity 0.90)
- Weakly-Supervised Action Localization by Hierarchically-Structured Latent Attention Modeling (ICCV 2023 · similarity 0.90)
- Learning to Localize Actions from Moments (ECCV 2020 · similarity 0.90)
- Temporal Action Localization in the Deep Learning Era: A Survey (TPAMI 2023 · similarity 0.90)
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.
