AI Paper Atlas

Research map / Video / Video-language understanding

ViLLa: Video Reasoning Segmentation with Large Language Model

Rongkun Zheng, Lu Qi, Xi Chen, Yi Wang, Kun Wang, Hengshuang Zhao

ICCV 2025 · Video › Video-language understanding · approach: Long video understanding · 40 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
add the context synthesizer (ablation)ViLLa baseline · VideoReasonSegaccuracy: 39.7 -> 43.3Improved
add the key segment extractor (ablation)ViLLa baseline · VideoReasonSegaccuracy: 39.7 -> 43.8Improved
add the hierarchical temporal synchronizer (ablation)ViLLa baseline · VideoReasonSegJ&F: 48.5 -> 50.7Improved
combine all three modules (ablation)ViLLa baseline · VideoReasonSegJ&F: 48.5 -> 55.4Improved
add multi-scale segmentation tokens only (without synchronization) (ablation)ViLLa · VideoReasonSegJ&F: 53.7 -> 54.2No clear effect
add video-frame synchronization on top of multi-scale tokens (ablation)ViLLa · VideoReasonSegJ&F: 54.2 -> 55.4Improved
use 4 aggregation layers rather than 3 (ablation)ViLLa · VideoReasonSegJ&F: 55.4 -> 55.1Hurt
use InternVideo2 as vision backbone instead of SAM (ablation)ViLLa · VideoReasonSegJ&F: 50.5 -> 55.4Improved
adaptive sampling (1/3 global, 2/3 neighbor) instead of global or neighbor only (ablation)Key segment extractor · VideoReasonSegJ&F: 54.7 -> 55.4Improved
use embedding-similarity aggregation instead of PixelLM-style feature fusion (ablation)ViLLa · Ref-DAVIS17J&F: 63.2 -> 64.4Improved
use ViLLa dedicated design instead of direct adaptation of LISA-style modules (ViLLa dagger) (ablation)ViLLa · VideoReasonSegJ&F: 51.9 -> 55.4Improved

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