English

Taming a Retrieval Framework to Read Images in Humanlike Manner for Augmenting Generation of MLLMs

Computer Vision and Pattern Recognition 2025-10-14 v1 Artificial Intelligence

Abstract

Multimodal large language models (MLLMs) often fail in fine-grained visual question answering, producing hallucinations about object identities, positions, and relations because textual queries are not explicitly anchored to visual referents. Retrieval-augmented generation (RAG) alleviates some errors, but it fails to align with human-like processing at both the retrieval and augmentation levels. Specifically, it focuses only on global-level image information but lacks local detail and limits reasoning about fine-grained interactions. To overcome this limitation, we present Human-Like Retrieval-Augmented Generation (HuLiRAG), a framework that stages multimodal reasoning as a ``what--where--reweight'' cascade. Queries are first anchored to candidate referents via open-vocabulary detection (what), then spatially resolved with SAM-derived masks to recover fine-grained precision (where), and adaptively prioritized through the trade-off between local and global alignment (reweight). Mask-guided fine-tuning further injects spatial evidence into the generation process, transforming grounding from a passive bias into an explicit constraint on answer formulation. Extensive experiments demonstrate that this human-like cascade improves grounding fidelity and factual consistency while reducing hallucinations, advancing multimodal question answering toward trustworthy reasoning.

Keywords

Cite

@article{arxiv.2510.10426,
  title  = {Taming a Retrieval Framework to Read Images in Humanlike Manner for Augmenting Generation of MLLMs},
  author = {Suyang Xi and Chenxi Yang and Hong Ding and Yiqing Ni and Catherine C. Liu and Yunhao Liu and Chengqi Zhang},
  journal= {arXiv preprint arXiv:2510.10426},
  year   = {2025}
}

Comments

12 pages, 5 figures