Research map / Video / Action recognition
Enriching Local and Global Contexts for Temporal Action Localization
Zixin Zhu, Wei Tang, Le Wang, Nanning Zheng, Gang Hua
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 L-Net before P-Net (non-local) (ablation) | P-Net (non-local) · THUMOS14 test | avg mAP: 39.76 -> 42.64 | Improved |
| add G-Net before P-Net (non-local) (ablation) | P-Net (non-local) · THUMOS14 test | avg mAP: 39.76 -> 41.33 | Improved |
| add L-Net before P-Net (P-GCN) (ablation) | P-Net (P-GCN) · THUMOS14 test | avg mAP: 42.43 -> 45.12 | Improved |
| add G-Net before P-Net (P-GCN) (ablation) | P-Net (P-GCN) · THUMOS14 test | avg mAP: 42.43 -> 43.51 | Improved |
| add L-Net before P-Net (P-GCN) (ablation) | P-Net (P-GCN) · THUMOS14 test (tIoU 0.5) | mAP: 49.10 -> 53.17 | Improved |
| add L-Net and G-Net to P-GCN (ablation) | P-Net (P-GCN) · THUMOS14 test (tIoU 0.5) | mAP: 49.10 -> 54.30 | Improved |
| add global context adaptation in G-Net (ablation) | ContextLoc · THUMOS14 test | avg mAP (fusion): 43.23 -> 45.70 | Improved |
| add interaction between adapted global contexts (ablation) | ContextLoc · THUMOS14 test | avg mAP (fusion): 44.14 -> 45.70 | Improved |
| process extended proposal as three proposals instead of one (dim 1024) (ablation) | ContextLoc · THUMOS14 test | avg mAP (fusion): 40.04 -> 45.70 | Improved |
| process extended proposal as three proposals instead of one (dim 3072) (ablation) | ContextLoc · THUMOS14 test | avg mAP (fusion): 44.57 -> 45.70 | Improved |
| add graph convolutional layers to P-GCN (Deep) (ablation) | P-GCN · THUMOS14 test | avg mAP: 42.43 -> 42.20 | No clear effect |
| add two graph convolutional layers to P-GCN (Deeper) (ablation) | P-GCN · THUMOS14 test | avg mAP: 42.43 -> 42.03 | Hurt |
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.
- G-TAD — G-TAD: Sub-Graph Localization for Temporal Action Detection (CVPR 2020) · better on 5 of 5 shared results
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
- ContextLoc++: A Unified Context Model for Temporal Action Localization (TPAMI 2023 · similarity 0.97)
- Graph Convolutional Module for Temporal Action Localization in Videos (TPAMI 2021 · similarity 0.92)
- ACSNet: Action-Context Separation Network for Weakly Supervised Temporal Action Localization (AAAI 2021 · similarity 0.92)
- Temporal Action Localization with Cross Layer Task Decoupling and Refinement (AAAI 2025 · similarity 0.91)
- Test-Time Zero-Shot Temporal Action Localization (CVPR 2024 · similarity 0.90)
- Bridge the Gap: From Weak to Full Supervision for Temporal Action Localization with PseudoFormer (CVPR 2025 · 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.
