Watch Wider and Think Deeper: Collaborative Cross-modal Chain-of-Thought for Complex Visual Reasoning

TL;DR

The paper introduces CoCoT, a novel multi-modal reasoning framework that dynamically grounds relevant image regions and enables multi-region collaboration, significantly improving complex visual reasoning accuracy.

Contribution

It proposes the CoCoT framework with dynamic multi-region grounding and relation-aware reasoning, along with a large-scale dataset for structured multi-region visual reasoning.

Findings

Achieves 15.4% accuracy improvement on LLaVA-1.5

Achieves 4.0% accuracy improvement on Qwen2-VL

Demonstrates effectiveness across six benchmarks

Abstract

Multi-modal reasoning requires the seamless integration of visual and linguistic cues, yet existing Chain-of-Thought methods suffer from two critical limitations in cross-modal scenarios: (1) over-reliance on single coarse-grained image regions, and (2) semantic fragmentation between successive reasoning steps. To address these issues, we propose the CoCoT (Collaborative Coross-modal Thought) framework, built upon two key innovations: a) Dynamic Multi-Region Grounding to adaptively detect the most relevant image regions based on the question, and b) Relation-Aware Reasoning to enable multi-region collaboration by iteratively aligning visual cues to form a coherent and logical chain of thought. Through this approach, we construct the CoCoT-70K dataset, comprising 74,691 high-quality samples with multi-region annotations and structured reasoning chains. Extensive experiments demonstrate that CoCoT significantly enhances complex visual reasoning, achieving an average accuracy improvement of 15.4% on LLaVA-1.5 and 4.0% on Qwen2-VL across six challenging benchmarks. The data and code are available at: https://github.com/deer-echo/CoCoT.