arXiv cs.AIMonday · May 25, 2026FREE

LFRAG: Layout-oriented Fine-grained Retrieval-Augmented Generation on Multimodal Document Understanding

LFRAG (Layout-oriented Fine-grained Retrieval-Augmented Generation) is a new framework from a paper on arXiv (2605.22829) that addresses limitations in existing multimodal RAG systems. Current systems rely on coarse-grained page-level retrieval, which fails to capture fine-grained semantic and layout structures in visually rich documents, leading to poor retrieval accuracy and redundant context. LFRAG advances multimodal RAG from page-level to block-level retrieval by performing layout segmentation to construct semantically coherent fine-grained retrieval units. It designs a semantic-layout fusion encoder that integrates local semantics with global context via cross-attention. With block-level late interaction retrieval, LFRAG enables precise query-content alignment and reduces irrelevant content for downstream generation. To enable rigorous evaluation, the authors constructed LFDocQA, a large-scale benchmark with block-level annotations spanning diverse document types. The paper is published on arXiv and was announced on May 25, 2026.