AI Paper Atlas

Research map / Video / Video-language understanding

VideoITG: Multimodal Video Understanding with Instructed Temporal Grounding

Shihao Wang, Guo Chen, De-An Huang, Zhiqi Li, Minghan Li, Guilin Liu, Jan Kautz, Jose M. Alvarez, Lei Zhang, Zhiding Yu

CVPR 2026 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 32 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 uniform sampling with VideoITG frame selectionLLaVA-OneVision-7B · LongVideoBench/MLVU/VideoMME avgaverage accuracy (%): 54.9 -> 61.6Improved
replace uniform sampling with VideoITG frame selectionQwen2-VL · LongVideoBench/MLVU/VideoMME avgaverage accuracy (%): 54.3 -> 60.3Improved
replace uniform 32 frames with VideoITG top-32six Video-LLMs (InternVL2.5/3.5, Qwen3-VL, LLaVA-Video, Eagle2.5) · LongVideoBench/MLVU/VideoMME/CG-Bench avgaverage accuracy (%): 58.7 -> 64.3 (InternVL2.5-8B); 61.8 -> 66.7 (Eagle2.5-8B)Improved
use VideoITG at 32 frames instead of uniform 64 framesLLaVA-Video-7B · LongVideoBench/MLVU/VideoMME avgaverage accuracy (%): 64.7 -> 66.9Improved
use VideoITG instead of QuoTA (both better than uniform)LLaVA-Video-7B · LongVideoBench/MLVU/VideoMME avgaverage accuracy (%): 65.9 -> 67.1Improved
use text-generation formulation instead of classification (ablation)VideoITG · VideoMME/MLVU/LongVideoBench avgaverage accuracy (%): 67.8 -> 48.5Hurt
use full bidirectional attention instead of causal+anchor (ablation)VideoITG · VideoMME/MLVU/LongVideoBench avgaverage accuracy (%): 67.2 -> 67.8Improved
scale grounding model from 3B to 7B (ablation)VideoITG · VideoMME/MLVU/LongVideoBench avgaverage accuracy (%): 66.8 -> 67.8Improved
add instructed clip captioning to annotation pipeline (ablation)VideoITG-40K · VideoMME Longaccuracy (%): 53.4 -> 56.9Improved
add instructed frame localization to annotation pipeline (ablation)VideoITG-40K · VideoMME Mediumaccuracy (%): 65.8 -> 67.1Improved
add video data to pre-training (ablation)VideoITG · VideoMME/MLVU/LongVideoBench avgaverage accuracy (%): 67.1 -> 67.8No clear effect
add image and video alignment pre-training (ablation)VideoITG · MLVUaccuracy (%): 69.1 -> 75.0Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most 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

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