Vision-centric retrieval for VQA requires retrieving images to supply missing visual cues and integrating them into the reasoning process. However, selecting the right images and integrating them effectively into the model's reasoning remains challenging.To address this challenge, we propose R3G, a modular Reasoning-Retrieval-Reranking framework.It first produces a brief reasoning plan that specifies the required visual cues, then adopts a two-stage strategy, with coarse retrieval followed by fine-grained reranking, to select evidence images.On MRAG-Bench, R3G improves accuracy across six MLLM backbones and nine sub-scenarios, achieving state-of-the-art overall performance. Ablations show that sufficiency-aware reranking and reasoning steps are complementary, helping the model both choose the right images and use them well. We release code and data at https://github.com/czh24/R3G.
@article{arxiv.2602.00104,
title = {R3G: A Reasoning--Retrieval--Reranking Framework for Vision-Centric Answer Generation},
author = {Zhuohong Chen and Zhengxian Wu and Zirui Liao and Shenao Jiang and Hangrui Xu and Yang Chen and Chaokui Su and Xiaoyu Liu and Haoqian Wang},
journal= {arXiv preprint arXiv:2602.00104},
year = {2026}
}