Research map / Video / Video-language understanding
An Efficient Framework for Dense Video Captioning
Maitreya Suin, A. N. Rajagopalan
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 |
|---|---|---|---|
| learned frame selection instead of uniform sampling (15 frames) (ablation) | RL frame-selection dense captioner (15 frames) · ActivityNet Captions val | METEOR: 5.44 -> 5.70 | Improved |
| learned frame selection instead of random sampling (15 frames) (ablation) | RL frame-selection dense captioner (15 frames) · ActivityNet Captions val | METEOR: 5.20 -> 5.70 | Improved |
| learned frame selection instead of uniform sampling (25 frames) (ablation) | RL frame-selection dense captioner (15 frames) · ActivityNet Captions val | BLEU-4: 1.42 -> 1.35 | Mixed |
| add knowledge distillation loss (ablation) | RL frame-selection dense captioner (15 frames) · ActivityNet Captions val | METEOR: 5.61 -> 5.70 | Improved |
| add global representation reward (ablation) | RL frame-selection dense captioner (15 frames) · ActivityNet Captions val | METEOR: 5.32 -> 5.70 | Improved |
| add event proposal reward (ablation) | RL frame-selection dense captioner (15 frames) · ActivityNet Captions val | METEOR: 5.07 -> 5.70 | Improved |
| add state encoder representation to proposal and caption modules (ablation) | RL frame-selection dense captioner (15 frames) · ActivityNet Captions val | METEOR: 5.14 -> 5.70 | Improved |
| add event, global and distillation terms beyond language reward only (ablation) | RL frame-selection dense captioner (15 frames) · ActivityNet Captions val | METEOR: 4.93 -> 5.70 | Improved |
| watch 25 frames instead of all frames (ablation) | RL frame-selection dense captioner (15 frames) · ActivityNet Captions val | METEOR: 6.21 -> 6.23 | No clear effect |
| watch 8 frames instead of 25 frames (ablation) | RL frame-selection dense captioner (15 frames) · ActivityNet Captions val | METEOR: 6.23 -> 4.98 | Hurt |
Closest papers by meaning
- Event-Equalized Dense Video Captioning (CVPR 2025 · similarity 0.91)
- Progress-Aware Video Frame Captioning (CVPR 2025 · similarity 0.89)
- Do You Remember? Dense Video Captioning with Cross-Modal Memory Retrieval (CVPR 2024 · similarity 0.89)
- Streaming Dense Video Captioning (CVPR 2024 · similarity 0.88)
- DIBS: Enhancing Dense Video Captioning with Unlabeled Videos via Pseudo Boundary Enrichment and Online Refinement (CVPR 2024 · similarity 0.88)
- Fine-Grained Captioning of Long Videos through Scene Graph Consolidation (ICML 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.
