Research map / Video / Video segmentation and tracking
You Can Ground Earlier Than See: An Effective and Efficient Pipeline for Temporal Sentence Grounding in Compressed Videos
Xiang Fang, Daizong Liu, Pan Zhou, Guoshun Nan
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 pseudo feature generation to baseline (ablation) | TCSF · ActivityNet Captions | R@1 IoU=0.3: 50.59 -> 60.25 | Improved |
| add three-branch spatial-temporal attention to baseline (ablation) | TCSF · ActivityNet Captions | R@1 IoU=0.3: 50.59 -> 62.79 | Improved |
| add adaptive motion-appearance fusion to baseline (ablation) | TCSF · ActivityNet Captions | R@1 IoU=0.3: 50.59 -> 63.74 | Improved |
| add adaptive fusion on top of PFG and TTA (ablation) | TCSF · ActivityNet Captions | R@1 IoU=0.3: 64.19 -> 66.87 | Improved |
| add motion vector and residual features to I-frame only (ablation) | TCSF · ActivityNet Captions | R@1 IoU=0.3: 64.27 -> 66.87 | Improved |
| use noun-formed query in spatial attention (ablation) | TCSF · ActivityNet Captions | R@1 IoU=0.3: 64.26 -> 66.87 | Improved |
| add temporal attention to spatial attention (ablation) | TCSF · ActivityNet Captions | R@1 IoU=0.5: 43.20 -> 48.38 | Improved |
| add pseudo appearance feature to pseudo motion feature (ablation) | TCSF · ActivityNet Captions | R@1 IoU=0.5: 47.51 -> 48.38 | Improved |
| increase temporal attention window K from 7 to 8 (ablation) | TCSF · ActivityNet Captions | R@1 IoU=0.3: 66.87 -> 65.92 | Hurt |
| use compressed-domain TCSF instead of fully decoded fully-supervised MMN | TCSF · Charades-STA | R@1 IoU=0.7: 27.28 -> 37.20 | Improved |
| use compressed-domain TCSF instead of fully decoded MMN | TCSF · ActivityNet Captions | R@1 IoU=0.5: 48.59 vs 48.38 | No clear effect |
| use compressed-domain TCSF instead of decoded-frame TAG for inference time | TCSF · ActivityNet Captions | total time of 100 videos (s): 243.36 -> 46.16 | Improved |
Closest papers by meaning
- Universal Video Temporal Grounding with Generative Multi-modal Large Language Models (NeurIPS 2025 · similarity 0.89)
- SnAG: Scalable and Accurate Video Grounding (CVPR 2024 · similarity 0.89)
- Empower Words: DualGround for Structured Phrase and Sentence-Level Temporal Grounding (NeurIPS 2025 · similarity 0.89)
- Temporal Sentence Grounding with Relevance Feedback in Videos (NeurIPS 2024 · similarity 0.89)
- Multi-Pair Temporal Sentence Grounding via Multi-Thread Knowledge Transfer Network (AAAI 2025 · similarity 0.89)
- Comprehensive Visual Grounding for Video Description (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.
