CoPRS: Learning Positional Prior from Chain-of-Thought for Reasoning Segmentation

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Lu, Zhenyu, Li, Liupeng, Wang, Jinpeng, Feng, Yan, Chen, Bin, Chen, Ke, Wang, Yaowei
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908901112283136
author Lu, Zhenyu
Li, Liupeng
Wang, Jinpeng
Feng, Yan
Chen, Bin
Chen, Ke
Wang, Yaowei
author_facet Lu, Zhenyu
Li, Liupeng
Wang, Jinpeng
Feng, Yan
Chen, Bin
Chen, Ke
Wang, Yaowei
contents Existing works on reasoning segmentation either connect hidden features from a language model directly to a mask decoder or represent positions in text, which limits interpretability and semantic detail. To solve this, we present CoPRS, a Multi-modal Chain-of-Thought (MCoT)-based positional perception model that bridges language reasoning to segmentation through a differentiable and interpretable positional prior instantiated as a heatmap. By making the reasoning process clear via MCoT and expressing it as a dense, differentiable heatmap, this interface enhances interpretability and diagnostic analysis and yields more concentrated evidence on the target. A learnable concentration token aggregates features of the image and reasoning text to generate this positional prior, which is decoded to precise masks through a lightweight decoder, providing a direct connection between reasoning and segmentation. Across the RefCOCO series and ReasonSeg, CoPRS matches or surpasses the best reported metrics on each standard split under comparable protocols, with performance at or above the prior state of the art across both validation and test partitions. Extensive experiments demonstrate a strong positive correlation among the CoT trajectory, the generated heatmap, and the decoded mask, supporting an interpretable alignment between the reasoning output and downstream mask generation. Collectively, these findings support the utility of this paradigm in bridging reasoning and segmentation and show advantages in concentration driven by reasoning and in more precise mask prediction. Code has been released at https://github.com/ZhenyuLU-Heliodore/CoPRS.
format Preprint
id arxiv_https___arxiv_org_abs_2510_11173
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CoPRS: Learning Positional Prior from Chain-of-Thought for Reasoning Segmentation
Lu, Zhenyu
Li, Liupeng
Wang, Jinpeng
Feng, Yan
Chen, Bin
Chen, Ke
Wang, Yaowei
Computer Vision and Pattern Recognition
Multimedia
Existing works on reasoning segmentation either connect hidden features from a language model directly to a mask decoder or represent positions in text, which limits interpretability and semantic detail. To solve this, we present CoPRS, a Multi-modal Chain-of-Thought (MCoT)-based positional perception model that bridges language reasoning to segmentation through a differentiable and interpretable positional prior instantiated as a heatmap. By making the reasoning process clear via MCoT and expressing it as a dense, differentiable heatmap, this interface enhances interpretability and diagnostic analysis and yields more concentrated evidence on the target. A learnable concentration token aggregates features of the image and reasoning text to generate this positional prior, which is decoded to precise masks through a lightweight decoder, providing a direct connection between reasoning and segmentation. Across the RefCOCO series and ReasonSeg, CoPRS matches or surpasses the best reported metrics on each standard split under comparable protocols, with performance at or above the prior state of the art across both validation and test partitions. Extensive experiments demonstrate a strong positive correlation among the CoT trajectory, the generated heatmap, and the decoded mask, supporting an interpretable alignment between the reasoning output and downstream mask generation. Collectively, these findings support the utility of this paradigm in bridging reasoning and segmentation and show advantages in concentration driven by reasoning and in more precise mask prediction. Code has been released at https://github.com/ZhenyuLU-Heliodore/CoPRS.
title CoPRS: Learning Positional Prior from Chain-of-Thought for Reasoning Segmentation
topic Computer Vision and Pattern Recognition
Multimedia
url https://arxiv.org/abs/2510.11173