Research map / Video / Action recognition
Semantic Proposal for Activity Localization in Videos via Sentence Query
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 |
|---|---|---|---|
| add proposal boundary refinement (ablation) | SAP · TACoS | R@1 IoU=0.5: 14.45 -> 18.24 | Helped |
| add proposal boundary refinement (ablation) | SAP · TACoS | R@1 IoU=0.1: 29.51 -> 31.15 | No clear effect |
| use Skip-thought instead of GloVe+LSTM sentence encoding (ablation) | SAP · TACoS | R@1 IoU=0.5: 16.62 -> 18.24 | Helped |
| use Skip-thought instead of GloVe+LSTM (ablation) | SAP · Charades-STA | R@1 IoU=0.7: 12.36 -> 13.36 | Helped |
| increase visual concepts from 93 to 912 (ablation) | SAP · TACoS | R@1 IoU=0.5: 16.60 -> 18.24 | Helped |
| increase visual concepts from 912 to 1413 (ablation) | SAP · TACoS | R@1 IoU=0.5: 18.24 -> 18.13 | Hurt |
| sentence-aware semantic proposals vs sliding-window proposals | CTRL · TACoS | R@1 IoU=0.5: 13.30 -> 18.24 | Helped |
| SAP vs ABLR | activity localization · Charades-STA | R@1 IoU=0.7: 9.01 -> 13.36 | Helped |
| fewer semantic proposals | proposal evaluation · TACoS | time per query: 1.76s -> 0.35s | Helped |
Closest papers by meaning
- Language-Driven Temporal Activity Localization: A Semantic Matching Reinforcement Learning Model (CVPR 2019 · similarity 0.91)
- To Find Where You Talk: Temporal Sentence Localization in Video with Attention Based Location Regression (AAAI 2019 · similarity 0.89)
- Boosting Weakly-Supervised Temporal Action Localization With Text Information (CVPR 2023 · similarity 0.89)
- Phrase-Level Temporal Relationship Mining for Temporal Sentence Localization (AAAI 2023 · similarity 0.88)
- HyrCap: Hybrid Rank-Calibration of Action Proposals for Temporal Event Understanding (NeurIPS 2026 · similarity 0.88)
- VideoGEM: Training-free Action Grounding in Videos (CVPR 2025 · similarity 0.86)
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.