MultiDocFusion: Hierarchical and Multimodal Chunking Pipeline for Enhanced RAG on Long Industrial Documents

TL;DR

MultiDocFusion introduces a hierarchical, multimodal chunking pipeline that significantly improves retrieval and QA performance on long industrial documents by leveraging document structure and multimodal information.

Contribution

It presents a novel multimodal chunking pipeline combining vision, OCR, and LLM-based hierarchical parsing to enhance RAG-based QA on complex industrial documents.

Findings

Improves retrieval precision by 8-15%.

Enhances QA scores (ANLS) by 2-3%.

Demonstrates the importance of document structure in QA.

Abstract

RAG-based QA has emerged as a powerful method for processing long industrial documents. However, conventional text chunking approaches often neglect complex and long industrial document structures, causing information loss and reduced answer quality. To address this, we introduce MultiDocFusion, a multimodal chunking pipeline that integrates: (i) detection of document regions using vision-based document parsing, (ii) text extraction from these regions via OCR, (iii) reconstruction of document structure into a hierarchical tree using large language model (LLM)-based document section hierarchical parsing (DSHP-LLM), and (iv) construction of hierarchical chunks through DFS-based grouping. Extensive experiments across industrial benchmarks demonstrate that MultiDocFusion improves retrieval precision by 8-15% and ANLS QA scores by 2-3% compared to baselines, emphasizing the critical role of explicitly leveraging document hierarchy for multimodal document-based QA. These significant performance gains underscore the necessity of structure-aware chunking in enhancing the fidelity of RAG-based QA systems.