AI Paper Atlas

Research map / Video / Video-language understanding

Mitigating Modality and Language-Style Gaps for Zero-Shot Video Moment Retrieval

Jihyun Lee, Cheol-Ho Cho, Woojin Jun, Woojin Jeong, Jae-Pil Heo

ECCV 2026 · Video › Video-language understanding · approach: Video moment retrieval · 1 citations (OpenAlex)

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 changedWhereMeasuredEffect
replace mean span scoring with top-k query-matching score (QMS) (ablation)Self-SiMS · QVHighlights valR1@0.5: 55.7 -> 59.5Improved
add self-matching span score (SMS) on QMS (ablation)Self-SiMS · QVHighlights valR1@0.5: 59.5 -> 60.3Improved
add MLLM yes/no re-ranking (ablation)Self-SiMS · QVHighlights valmAP@avg: 38.7 -> 39.3Improved
generate spans from self-similarity instead of sliding window (ablation)Self-SiMS · QVHighlights valmAP@avg: 12.5 -> 39.3Improved
generate spans from self-similarity instead of k-means clustering (ablation)Self-SiMS · QVHighlights valmAP@avg: 31.3 -> 39.3Improved
use DINOv2 instead of CLIP-B/32 frame features for self-similarity (ablation)Self-SiMS · QVHighlights valmAP@avg: 35.3 -> 38.7Improved
add SMS across CLIP-B/32 encoder (ablation)Self-SiMS · QVHighlights valmAP@avg: 31.5 -> 35.3Improved
use larger text encoder (E5-Mistral, NV-Embed) for query-caption similarity (ablation)VTG-GPT · QVHighlights valmAP@avg: 34.9 -> 30.87 / 31.44No clear effect
fuse frame and caption self-similarity maps versus frame-only (before re-ranking) (ablation)Self-SiMS · QVHighlights valmAP@avg: 38.2 -> 38.7Improved
swap LLaMA-3.2-Vision for MiniGPT-v2 in captioning and re-ranking (ablation)Self-SiMS · QVHighlights valmAP@avg: 39.3 -> 40.0No clear effect
add SMS to QMS on severe language-style gap subset (ablation)Self-SiMS · QVHighlights val (severe LS-gap subset)R1@0.5: 31.3 -> 31.7Improved
dynamic instance-wise threshold versus fixed thresholds (ablation)Self-SiMS · three datasetsmIoUImproved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy