Research map / Video / Video-language understanding
Sketch, Ground, and Refine: Top-Down Dense Video Captioning
Chaorui Deng, Shizhe Chen, Da Chen, Yuan He, Qi Wu
Check 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 |
|---|---|---|---|
| generate a video-level paragraph first, then ground and refine its event sentences (ablation) | SGR · ActivityNet Captioning validation | METEOR: 7.79 -> 9.07 | Improved |
| generate a video-level paragraph first, then ground and refine its event sentences (ablation) | SGR · YouCook2 | METEOR: 3.67 -> 4.35 | Improved |
| add caption refinement without a second grounding stage (ablation) | SG · ActivityNet Captioning validation (C3D) | METEOR: 7.79 -> 8.85 | Improved |
| add a second grounding stage using the refined captions (ablation) | SG with refine caption · ActivityNet Captioning validation (C3D) | METEOR: 8.85 -> 9.07 | Improved |
| add caption refinement without a second grounding stage (ablation) | SG · ActivityNet Captioning validation (TSN) | METEOR: 8.27 -> 9.13 | Improved |
| add a second grounding stage using the refined captions (ablation) | SG with refine caption · ActivityNet Captioning validation (TSN) | METEOR: 9.13 -> 9.37 | Improved |
| add caption refinement without a second grounding stage (ablation) | SG · YouCook2 | METEOR: 3.67 -> 3.96 | Improved |
| add a second grounding stage using the refined captions (ablation) | SG with refine caption · YouCook2 | METEOR: 3.96 -> 4.35 | Improved |
| use the sentence-level advantage in RL refinement rewards (ablation) | SGR refinement training · ActivityNet Captioning | METEOR | Improved |
| use the paragraph-level advantage in RL refinement rewards (ablation) | SGR refinement training · ActivityNet Captioning | METEOR | Improved |
| use both hierarchical advantage functions instead of cross-entropy refinement alone (ablation) | SGR refinement training · ActivityNet Captioning | METEOR | Improved |
| attend to the aligned video segment during caption refinement (ablation) | SGR dual-path cross attention · ActivityNet Captioning | METEOR | Improved |
Closest papers by meaning
- Streamlined Dense Video Captioning (CVPR 2019 · similarity 0.94)
- Event-Equalized Dense Video Captioning (CVPR 2025 · similarity 0.92)
- Progress-Aware Video Frame Captioning (CVPR 2025 · similarity 0.89)
- Exploring Temporal Event Cues for Dense Video Captioning in Cyclic Co-Learning (AAAI 2025 · similarity 0.89)
- Dense Video Object Captioning from Disjoint Supervision (ICLR 2025 · similarity 0.89)
- AuroraCap: Efficient, Performant Video Detailed Captioning and a New Benchmark (ICLR 2025 · 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.
