ConceptSeg-R1: Segment Any Concept via Meta-Reinforcement Learning

arXiv:2605.20385 · 2026-05-21 공개 · arXiv · PDF

medical-imaging promptable-segmentation industrial-applications meta-reinforcement-learning rule-induced-grounding context-reasoning concept-taxonomy shortcut-routing

Abstract

Recent progress in promptable segmentation has shifted visual perception from object-level localization toward concept-level understanding. However, the notion of a concept remains under-specified, making it unclear whether current methods truly generalize beyond category recognition. In this work, we formalize generalized concept segmentation through a three-level taxonomy consisting of context-independent (CI), context-dependent (CD), and context-reasoning (CR) concepts, which reveals a clear capability gap across increasing levels of cognitive complexity. To address this challenge, we propose ConceptSeg-R1, a unified framework that reformulates concept segmentation as rule-induced concept grounding. At the core of our method is Meta-GRPO, a meta-reinforcement learning mechanism that learns transferable task rules from visual demonstrations and verifies them through proxy reasoning. The inferred reasoning states are then translated into segmentation-ready concept prompts via a lightweight concept translation module, enabling deductive application to target images. A shortcut routing strategy further preserves the native efficiency of segmentation models on simple cases. To systematically evaluate generalized concept segmentation, we conduct extensive experiments across diverse CI, CD, and CR concept segmentation benchmarks spanning natural, industrial, medical and reasoning-intensive domains. Without bells and whistles, ConceptSeg-R1 achieves strong performance across the full concept hierarchy while maintaining the native capability of promptable segmentation backbones. As an initial step toward segmenting any concept, we hope ConceptSeg-R1 can serve as a practical baseline for advancing segmentation from object-level prediction toward concept-level understanding.

한국어 요약

📋 한 줄 요약

**[프롬프트 분할 / 메타강화학습]** 일반화된 concept segmentation을 CI/CD/CR 3단 분류로 정의하고 메타 GRPO로 transferable한 규칙을 학습하는 통합 프레임워크 ConceptSeg-R1 제안.

🎯 핵심 기여도

💡 핵심 아이디어

"개념"은 고정된 카테고리 사전이 아니라 맥락·추론에 따라 변하는 동적 개체이며, "전이 가능한 규칙"을 학습 단위로 두고 메타 강화학습으로 일반화하면 분할 모델이 객체 인식 너머로 확장될 수 있다.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: "객체 분할 → 개념 분할"이라는 segmentation 패러다임 전환의 실용 베이스라인 제공. **한계**: rule 학습이 시각 데모 품질에 의존, 매우 모호하거나 문화적 개념에 대한 처리는 본문 범위 밖.

🚀 실용적 활용