CGC: Compositional Grounded Contrast for Fine-Grained Multi-Image Understanding

TL;DR

This paper introduces CGC, a cost-effective framework that enhances fine-grained multi-image understanding in multimodal models through compositional contrast and spatial reasoning, achieving state-of-the-art results.

Contribution

CGC leverages existing annotations to construct multi-image training instances with contrastive learning and rule-based spatial rewards, improving fine-grained understanding without extensive annotations.

Findings

CGC achieves state-of-the-art on MIG-Bench and VLM2-Bench.

It improves model performance on multiple multimodal reasoning tasks.

The framework enhances object constancy and spatial alignment in multi-image understanding.

Abstract

Although Multimodal Large Language Models (MLLMs) have advanced rapidly, they still face notable challenges in fine-grained multi-image understanding, often exhibiting spatial hallucination, attention leakage, and failures in object constancy. In addition, existing approaches typically rely on expensive human annotations or large-scale chain-of-thought (CoT) data generation. We propose Compositional Grounded Contrast (abbr. CGC), a low-cost full framework for boosting fine-grained multi-image understanding of MLLMs. Built on existing single-image grounding annotations, CGC constructs compositional multi-image training instances through Inter-Image Contrast and Intra-Image Contrast, which introduce semantically decoupled distractor contexts for cross-image discrimination and correlated cross-view samples for object constancy, respectively. CGC further introduces a Rule-Based Spatial Reward within the GRPO framework to improve source-image attribution, spatial alignment, and structured output validity under a Think-before-Grounding paradigm. Experiments show that CGC achieves state-of-the-art results on fine-grained multi-image benchmarks, including MIG-Bench and VLM2-Bench. The learned multi-image understanding capability also transfers to broader multimodal understanding and reasoning tasks, yielding consistent gains over the Qwen3-VL-8B base model on MathVista (+2.90), MuirBench (+2.88), MMStar (+1.93), MMMU (+1.77), and BLINK (+1.69).