AI Paper Atlas

Research map / Video / Video-language understanding

VideoSeg-R1:Reasoning Video Object Segmentation via Reinforcement Learning

Zishan Xu, Yifu Guo, Yuquan Lu, Fengyu Yang, Junxin Li, Lihua Cai

AAAI 2026 · Video › Video-language understanding · approach: Long video understanding · 4 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
train with RL instead of SFT (ablation)VideoSeg-R1 (Qwen2.5-VL-7B) · ReVOSJ&F: 50.9 -> 58.9Improved
train with RL instead of SFT (ablation)VideoSeg-R1 · ReasonSeggIoU: 59.4 -> 67.2Improved
add CoT reasoning reward to RL (ablation)VideoSeg-R1 · ReVOSJ&F: 58.9 -> 61.1Improved
add CoT reasoning reward to RL (ablation)VideoSeg-R1 · ReasonSeggIoU: 67.2 -> 69.8Improved
use hierarchical key-segment + target-frame sampler instead of first frame (ablation)VideoSeg-R1 · MeViSJ&F: 45.7 -> 55.3Improved
add key-segment stage on top of target-frame localization (ablation)VideoSeg-R1 · ReVOSJ&F: 57.6 -> 61.1Improved
use random frame from key segment instead of target-frame localization (ablation)VideoSeg-R1 · MeViSJ&F: 55.3 -> 49.2Hurt
add bounding box prompt over no-prompt baseline (ablation)VideoSeg-R1 · ReVOSJ&F: 40.2 -> 57.2Improved
add central point to Bbox (ablation)VideoSeg-R1 · ReasonSeggIoU: 65.7 -> 68.2Improved
add negative point to Bbox + Pcentral (ablation)VideoSeg-R1 · ReVOSJ&F: 60.3 -> 61.1Improved
add soft length penalty (ablation)VideoSeg-R1 · Ref-DAVIS17J&F: 75.2 -> 79.8Improved
add soft length penalty (ablation)VideoSeg-R1 · ReVOSreasoning tokens: 89.9 -> 56.1Improved

Later papers that compared with this method

1 later accepted paper reports VideoSeg-R1 as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

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