Reinforcing Video Reasoning Segmentation to Think Before It Segments

TL;DR

This paper introduces Veason-R1, a reinforcement learning-based LVLM designed for video reasoning segmentation, which improves interpretability, spatiotemporal reasoning, and achieves state-of-the-art results on multiple benchmarks.

Contribution

The paper presents Veason-R1, a novel LVLM for VRS that incorporates structured reasoning via Chain-of-Thought training and Group Relative Policy Optimization, enhancing performance and interpretability.

Findings

Achieves +1.3 J &F on ReVOS and +10.0 J &F on ReasonVOS benchmarks.

Demonstrates robustness to hallucinations with +8.8 R improvement.

Outperforms prior methods significantly in video reasoning segmentation.

Abstract

Video reasoning segmentation (VRS) endeavors to delineate referred objects in videos guided by implicit instructions that encapsulate human intent and temporal logic. Previous approaches leverage large vision language models (LVLMs) to encode object semantics into <SEG> tokens for mask prediction. However, this paradigm suffers from limited interpretability during inference and suboptimal performance due to inadequate spatiotemporal reasoning. Drawing inspiration from seminal breakthroughs in reinforcement learning, we introduce Veason-R1, a specialized LVLM for VRS that emphasizes structured reasoning in segmentation. Veason-R1 is trained through Group Relative Policy Optimization (GRPO) augmented with Chain-of-Thought (CoT) initialization. To begin with, we curate high-quality CoT training data to instill structured reasoning trajectories, bridging video-level semantics and frame-level spatial grounding, yielding the supervised fine-tuned model Veason-SFT. Subsequently, GRPO fine-tuning encourages efficient exploration of the reasoning space by optimizing reasoning chains. To this end, we incorporate a holistic reward mechanism that synergistically enhances spatial alignment and temporal consistency, bolstering keyframe localization and fine-grained grounding. Comprehensive empirical evaluations demonstrate that Veason-R1 achieves state-of-the-art performance on multiple benchmarks, surpassing prior art by significant margins (e.g., +1.3 J &F in ReVOS and +10.0 J &F in ReasonVOS), while exhibiting robustness to hallucinations (+8.8 R). Our code and model weights will be available at Veason-R1.