MIRROR: Multimodal Iterative Reasoning via Reflection on Visual Regions

TL;DR

MIRROR introduces a multimodal iterative reasoning framework that employs visual reflection and verification to improve reasoning accuracy and reduce hallucinations in vision-language models.

Contribution

The paper presents MIRROR, a novel reflection-based iterative reasoning framework with a new dataset, ReflectV, enabling models to verify and revise answers grounded in visual evidence.

Findings

MIRROR improves reasoning correctness on benchmarks.

Reduces visual hallucinations in model outputs.

Enhances evidence-based verification in multimodal reasoning.

Abstract

In the era of Vision-Language Models (VLMs), enhancing multimodal reasoning capabilities remains a critical challenge, particularly in handling ambiguous or complex visual inputs, where initial inferences often lead to hallucinations or logic errors. Existing VLMs often produce plausible yet ungrounded answers, and even when prompted to "reflect", their corrections may remain detached from the image evidence. To address this, we propose the MIRROR framework for Multimodal Iterative Reasoning via Reflection On visual Regions. By embedding visual reflection as a core mechanism, MIRROR is formulated as a closed-loop process comprising draft, critique, region-based verification, and revision, which are repeated until the output is visually grounded. To facilitate training of this model, we construct **ReflectV**, a visual reflective dataset for multi-turn supervision that explicitly contains reflection triggers, region-based verification actions, and answer revision grounded in visual evidence. Experiments on both general vision-language benchmarks and representative vision-language reasoning benchmarks show that MIRROR improves correctness and reduces visual hallucinations, demonstrating the value of training reflection as an evidence-seeking, region-aware verification process rather than a purely textual revision step.