Research map / Video / Video-language understanding
See, Rank, and Filter: Important Word-Aware Clip Filtering via Scene Understanding for Moment Retrieval and Highlight Detection
YuEun Lee, Jung Uk 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 MLLM captions (InternVL2) to baseline (ablation) | baseline detector · QVHighlights val | HD HIT@1: 64.77 -> 67.94 | Improved |
| add RFM (iterative filtering) to FEM+captions (ablation) | proposed method · QVHighlights val | MR R1@0.7: 53.81 -> 55.68 | Improved |
| add FEM to captions+RFM (ablation) | proposed method · QVHighlights val | HD HIT@1: 68.00 -> 71.23 | Improved |
| add FEM alone to baseline (ablation) | baseline detector · QVHighlights val | MR R1@0.7: 49.74 -> 49.81 | No clear effect |
| add RFM alone to baseline (ablation) | baseline detector · QVHighlights val | MR R1@0.7: 49.74 -> 49.61 | No clear effect |
| add MLLM captions to prior SOTA models (ablation) | TR-DETR, UVCOM, Keyword-DETR · QVHighlights val | HD HIT@1: 68.13-69.55 vs ours 71.23 | No clear effect |
| RFM filtering iterations 0 -> 5 (ablation) | RFM · QVHighlights val | MR R1@0.7: 53.81 -> 55.68 | Improved |
| RFM filtering iterations 5 -> 7 (ablation) | RFM · QVHighlights val | HD HIT@1: 71.23 -> 69.29 | Hurt |
| use LLaVA instead of InternVL2 for captions (ablation) | proposed method · QVHighlights val | HD HIT@1: 71.23 -> 69.68 | Mixed |
| proposed method vs Keyword-DETR in audio+video setting | proposed method · QVHighlights test (video + audio) | MR mAP@0.5: 68.30 -> 68.09 | Mixed |
| proposed method vs Keyword-DETR on Charades-STA SF+C | proposed method · Charades-STA (SlowFast+CLIP features) | R1@0.5: 61.08 -> 60.97 | Mixed |
Closest papers by meaning
- Watch Video, Catch Keyword: Context-aware Keyword Attention for Moment Retrieval and Highlight Detection (AAAI 2025 · similarity 0.90)
- Fewer Steps, Better Performance: Efficient Cross-Modal Clip Trimming for Video Moment Retrieval Using Language (AAAI 2024 · similarity 0.89)
- Bridging the Gap: A Unified Video Comprehension Framework for Moment Retrieval and Highlight Detection (CVPR 2024 · similarity 0.89)
- TR-DETR: Task-Reciprocal Transformer for Joint Moment Retrieval and Highlight Detection (AAAI 2024 · similarity 0.89)
- ShotVL: Human-Centric Highlight Frame Retrieval via Language Queries (AAAI 2025 · similarity 0.87)
- Generative Video Diffusion for Unseen Novel Semantic Video Moment Retrieval (AAAI 2025 · similarity 0.87)
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.
