[4C2-OS-27b] [Organized Session] OS-27

[4C2-OS-27b-01] Medical Image Report Findings Structurization

〇JIANWEI XIE¹, Fujio Toriumi¹ (1. The University of Tokyo)

Keywords:NLP, Text Mining, Medical Text Processing

Structuring of a medical report has various applications. But most of them are only for supporting professionals to write well-structured reports, which means it cannot structurize reports without manual checks. In this study, we propose an automatic method to structurize medical image report findings. It extracts important information like disease names or symptoms in sentences and marks them with proper tags. With these tags, professionals can easily get the information they need from the report. Furthermore, we apply the method to a real dataset and get remarkably good results.