CoReS: Orchestrating the Dance of Reasoning and Segmentation
This work addresses the problem of accurate object localization in complex reasoning scenarios for computer vision and AI researchers, representing an incremental advance in method design.
The paper tackles the reasoning segmentation task, where Multi-modal Large Language Models struggle to localize objects in complex reasoning contexts, by introducing CoReS, a dual-chain structure that mimics human visual search to enhance segmentation, achieving a 6.5% improvement over state-of-the-art on the ReasonSeg dataset.
The reasoning segmentation task, which demands a nuanced comprehension of intricate queries to accurately pinpoint object regions, is attracting increasing attention. However, Multi-modal Large Language Models (MLLM) often find it difficult to accurately localize the objects described in complex reasoning contexts. We believe that the act of reasoning segmentation should mirror the cognitive stages of human visual search, where each step is a progressive refinement of thought toward the final object. Thus we introduce the Chains of Reasoning and Segmenting (CoReS) and find this top-down visual hierarchy indeed enhances the visual search process. Specifically, we propose a dual-chain structure that generates multi-modal, chain-like outputs to aid the segmentation process. Furthermore, to steer the MLLM's outputs into this intended hierarchy, we incorporate in-context inputs as guidance. Extensive experiments demonstrate the superior performance of our CoReS, which surpasses the state-of-the-art method by 6.5\% on the ReasonSeg dataset. Project: https://chain-of-reasoning-and-segmentation.github.io/.