A Graphical Approach to Document Layout Analysis

TL;DR

This paper introduces GLAM, a lightweight graph neural network for document layout analysis that leverages PDF metadata, achieving competitive accuracy with significantly reduced model size and improved efficiency.

Contribution

The paper presents GLAM, a novel graph neural network model that outperforms larger vision-based models on DLA tasks by utilizing structured PDF metadata.

Findings

GLAM outperforms a 140M+ parameter model on several classes.

Ensemble of GLAM and vision model achieves new state-of-the-art.

GLAM is over 5 times more efficient than existing models.

Abstract

Document layout analysis (DLA) is the task of detecting the distinct, semantic content within a document and correctly classifying these items into an appropriate category (e.g., text, title, figure). DLA pipelines enable users to convert documents into structured machine-readable formats that can then be used for many useful downstream tasks. Most existing state-of-the-art (SOTA) DLA models represent documents as images, discarding the rich metadata available in electronically generated PDFs. Directly leveraging this metadata, we represent each PDF page as a structured graph and frame the DLA problem as a graph segmentation and classification problem. We introduce the Graph-based Layout Analysis Model (GLAM), a lightweight graph neural network competitive with SOTA models on two challenging DLA datasets - while being an order of magnitude smaller than existing models. In particular, the 4-million parameter GLAM model outperforms the leading 140M+ parameter computer vision-based model on 5 of the 11 classes on the DocLayNet dataset. A simple ensemble of these two models achieves a new state-of-the-art on DocLayNet, increasing mAP from 76.8 to 80.8. Overall, GLAM is over 5 times more efficient than SOTA models, making GLAM a favorable engineering choice for DLA tasks.