Research map / Video / Video-language understanding
GranAlign: Granularity-Aware Alignment Framework for Zero-shot Video Moment Retrieval
Mingyu Jeon, Sunjae Yoon, Jonghee Kim, Junyeong 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 |
|---|---|---|---|
| use query-aware caption instead of query-agnostic caption with raw query (ablation) | GranAlign · QVHighlights val | mAP@avg: 31.8 -> 32.13 | Improved |
| pair simplified query with query-aware caption (granularity mismatch) (ablation) | GranAlign · QVHighlights val | mAP@avg: 37.13 -> 36.22 | Hurt |
| pair detailed query with query-aware caption instead of query-agnostic (ablation) | GranAlign · QVHighlights val | mAP@avg: 36.82 -> 37.65 | Improved |
| combine simplified-agnostic and detailed-aware pairs versus best single pair (ablation) | GranAlign · QVHighlights val | mAP@avg: 37.65 -> 39.12 | Improved |
| combine both pairs versus best single pair (ablation) | GranAlign · QVHighlights val | R1@0.5: 59.48 -> 61.94 | Improved |
| use simplified query with query-agnostic caption instead of raw query (ablation) | GranAlign · QVHighlights val | mAP@avg: 31.8 -> 37.13 | Improved |
| swap granularity assignment (detailed with agnostic, simplified with aware) (ablation) | GranAlign · QVHighlights val | R1@0.5: 61.94 -> 60.90 | Hurt |
| use detailed queries for both caption types instead of granularity-matched pairing (ablation) | GranAlign · QVHighlights val | mAP@avg: 39.12 -> 39.01 | Hurt |
| use three rewritten queries and weight 0.3 versus nearby values (ablation) | GranAlign · QVHighlights val | performance | No clear effect |
| GranAlign versus Moment-GPT on QVHighlights val | GranAlign · QVHighlights val | mAP@avg: 35.9 -> 39.12 | Improved |
| GranAlign versus Moment-GPT on ActivityNet-Captions | GranAlign · ActivityNet-Captions | R1@0.5: 31.1 -> 34.0 | Improved |
| GranAlign versus Moment-GPT on Charades-STA | GranAlign · Charades-STA | R1@0.5: 38.4 -> 39.6 | Improved |
Closest papers by meaning
- Zero-shot Video Moment Retrieval via Off-the-shelf Multimodal Large Language Models (AAAI 2025 · similarity 0.90)
- Towards Balanced Alignment: Modal-Enhanced Semantic Modeling for Video Moment Retrieval (AAAI 2024 · similarity 0.90)
- Generative Video Diffusion for Unseen Novel Semantic Video Moment Retrieval (AAAI 2025 · similarity 0.89)
- Anchor-Aware Similarity Cohesion in Target Frames Enables Predicting Temporal Moment Boundaries in 2D (CVPR 2025 · similarity 0.89)
- Bridging the Semantic Granularity Gap Between Text and Frame Representations for Partially Relevant Video Retrieval (AAAI 2025 · similarity 0.89)
- VERIFIED: A Video Corpus Moment Retrieval Benchmark for Fine-Grained Video Understanding (NeurIPS 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.
