{"work":{"id":"7a33b2a4-8409-4a00-ad7b-84e810df1ba7","openalex_id":null,"doi":null,"arxiv_id":"2503.06520","raw_key":null,"title":"Seg-Zero: Reasoning-Chain Guided Segmentation via Cognitive Reinforcement","authors":null,"authors_text":"Yuqi Liu, Bohao Peng, Zhisheng Zhong, Zihao Yue, Fanbin Lu, Bei Yu","year":2025,"venue":"cs.CV","abstract":"Traditional methods for reasoning segmentation rely on supervised fine-tuning with categorical labels and simple descriptions, limiting its out-of-domain generalization and lacking explicit reasoning processes. To address these limitations, we propose Seg-Zero, a novel framework that demonstrates remarkable generalizability and derives explicit chain-of-thought reasoning through cognitive reinforcement. Seg-Zero introduces a decoupled architecture consisting of a reasoning model and a segmentation model. The reasoning model interprets user intentions, generates explicit reasoning chains, and produces positional prompts, which are subsequently used by the segmentation model to generate precious pixel-level masks. We design a sophisticated reward mechanism that integrates both format and accuracy rewards to effectively guide optimization directions. Trained exclusively via reinforcement learning with GRPO and without explicit reasoning data, Seg-Zero achieves robust zero-shot generalization and exhibits emergent test-time reasoning capabilities. Experiments show that Seg-Zero-7B achieves a zero-shot performance of 57.5 on the ReasonSeg benchmark, surpassing the prior LISA-7B by 18\\%. This significant improvement highlights Seg-Zero's ability to generalize across domains while presenting an explicit reasoning process.","external_url":"https://arxiv.org/abs/2503.06520","cited_by_count":null,"metadata_source":"pith","metadata_fetched_at":"2026-07-04T19:30:08.050016+00:00","pith_arxiv_id":"2503.06520","created_at":"2026-05-09T05:45:22.055978+00:00","updated_at":"2026-07-04T19:30:08.050016+00:00","title_quality_ok":true,"display_title":"Seg-Zero: Reasoning-Chain Guided Segmentation via Cognitive Reinforcement","render_title":"Seg-Zero: Reasoning-Chain Guided Segmentation via Cognitive Reinforcement"},"hub":{"state":{"work_id":"7a33b2a4-8409-4a00-ad7b-84e810df1ba7","tier":"hub","tier_reason":"10+ Pith inbound or 1,000+ external citations","pith_inbound_count":53,"external_cited_by_count":null,"distinct_field_count":4,"first_pith_cited_at":"2025-02-24T18:50:52+00:00","last_pith_cited_at":"2026-07-02T17:56:49+00:00","author_build_status":"not_needed","summary_status":"needed","contexts_status":"needed","graph_status":"needed","ask_index_status":"not_needed","reader_status":"not_needed","recognition_status":"not_needed","updated_at":"2026-08-22T07:29:36.474455+00:00","tier_text":"hub"},"tier":"hub","role_counts":[{"context_role":"background","n":7},{"context_role":"baseline","n":2}],"polarity_counts":[{"context_polarity":"background","n":7},{"context_polarity":"baseline","n":2}],"runs":{},"summary":{},"graph":{},"authors":[]}}