Document Information Extraction via Global Tagging

Shaojie He, Tianshu Wang, Yaojie Lu, Hongyu Lin, Xianpei Han, Yingfei Sun, Le Sun

Abstract: Document Information Extraction (DIE) is a crucial task for extracting key information from visually-rich documents. The typical pipeline approach for this task involves Optical Character Recognition (OCR), serializer, Semantic Entity Recognition (SER), and Relation Extraction (RE) modules. However, this pipeline presents significant challenges in real-world scenarios due to issues such as unnatural text order and error propagation between different modules. To address these challenges, we propose a novel tagging-based method – Global TaggeR (GTR), which converts the original sequence labeling task into a token relation classification task. This approach globally links discontinuous semantic entities in complex layouts, and jointly extracts entities and relations from documents. In addition, we design a joint training loss and a joint decoding strategy for SER and RE tasks based on GTR. Our experiments on multiple datasets demonstrate that GTR not only mitigates the issue of text in the wrong order but also improves RE performance.