Research map / Video / Video-language understanding
Follow the Saliency: Supervised Saliency for Retrieval-augmented Dense Video Captioning
Seung hee Choi, MinJu Jeon, Hyunwoo Oh, Jihwan Lee, Dong-Jin Kim
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 SGSR (ablation) | Vid2Seq baseline · YouCook2 (val) | CIDEr: 66.29 -> 76.94 | Improved |
| add SaliP (ablation) | Vid2Seq baseline · YouCook2 (val) | CIDEr: 66.29 -> 78.74 | Improved |
| add SaliP on top of SGSR (ablation) | STaRC · YouCook2 (val) | CIDEr: 76.94 -> 80.53 | Improved |
| add SWSA feature refinement (ablation) | STaRC · YouCook2 (val) | CIDEr: 75.82 -> 80.53 | Improved |
| use window sizes 8,32,64 instead of 10,40,70 (ablation) | SWSA · YouCook2 (val) | CIDEr: 70.52 -> 80.53 | Improved |
| use SGSR instead of k-means (ablation) | STaRC · YouCook2 (val) | CIDEr: 75.63 -> 80.53 | Improved |
| use SGSR instead of Sali4Vid adaptive clustering (ablation) | STaRC · YouCook2 (val) | CIDEr: 78.19 -> 80.53 | Improved |
| saliency-weighted pooling instead of uniform average (ablation) | SGSR · YouCook2 (val) | CIDEr: 78.95 -> 80.53 | Improved |
| retrieve 10 captions instead of 15 (ablation) | STaRC · YouCook2 (val) | CIDEr: 78.00 -> 80.53 | Improved |
| retrieve 10 captions instead of 3 (ablation) | STaRC · YouCook2 (val) | CIDEr: 79.65 -> 80.53 | Improved |
| replace accurate saliency prompt with Gaussian noise prompt (ablation) | SaliP · YouCook2 (val) | CIDEr: 80.53 -> 73.47 | Hurt |
| replace saliency prompt with zero prompt (ablation) | SaliP · YouCook2 (val) | CIDEr: 80.53 -> 76.84 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- HiCM2 — HiCM²: Hierarchical Compact Memory Modeling for Dense Video Captioning (AAAI 2025) · better on 4 of 4 shared results
- Vid2Seq — Vid2Seq: Large-Scale Pretraining of a Visual Language Model for Dense Video Captioning (CVPR 2023) · better on 4 of 4 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
- HiCM²: Hierarchical Compact Memory Modeling for Dense Video Captioning (AAAI 2025 · similarity 0.87)
- DIBS: Enhancing Dense Video Captioning with Unlabeled Videos via Pseudo Boundary Enrichment and Online Refinement (CVPR 2024 · similarity 0.87)
- SALOVA: Segment-Augmented Long Video Assistant for Targeted Retrieval and Routing in Long-Form Video Analysis (CVPR 2025 · similarity 0.86)
- Event-Equalized Dense Video Captioning (CVPR 2025 · similarity 0.86)
- Dense Video Object Captioning from Disjoint Supervision (ICLR 2025 · similarity 0.86)
- Exploiting Auxiliary Caption for Video Grounding (AAAI 2024 · 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.
