DocFusion: A Unified Framework for Document Parsing Tasks

TL;DR

DocFusion is a lightweight, unified generative model that efficiently handles multiple document parsing tasks, achieving state-of-the-art results by leveraging task interactions and integrated training.

Contribution

It introduces a novel unified framework for document parsing that simplifies architecture and improves performance through collaborative training and task interaction modeling.

Findings

Achieves SOTA performance across four document parsing tasks

Leverages mutual benefits among recognition tasks

Significantly improves detection accuracy with integrated data

Abstract

Document parsing is essential for analyzing complex document structures and extracting fine-grained information, supporting numerous downstream applications. However, existing methods often require integrating multiple independent models to handle various parsing tasks, leading to high complexity and maintenance overhead. To address this, we propose DocFusion, a lightweight generative model with only 0.28B parameters. It unifies task representations and achieves collaborative training through an improved objective function. Experiments reveal and leverage the mutually beneficial interaction among recognition tasks, and integrating recognition data significantly enhances detection performance. The final results demonstrate that DocFusion achieves state-of-the-art (SOTA) performance across four key tasks.