[4Rin1] Interactive 2

[4Rin1-52] Research on more efficient input of prescription information using OCR and LINE interface

〇Taichi Obara¹, Kenji Tanaka¹, Yuki Matsuda¹ (1.Systems Innovation, Faculty of Engineering, The University of Tokyo)

Keywords:OCR, Image recognition, machine learning, Character recognition

As society as a whole demands automation and IT, the application field of machine learning technology is also expanding. The purpose of this research is to improve operational efficiency in the medical field, where automation and IT have not yet progressed, and to achieve this, OCR technology was used. Prescription information is now manually entered into computers at many pharmacies. Therefore, we proposed a system that uses LINE as an interface to obtain information from prescription images taken with a smartphone using OCR. The preprocessing and text processing methods proposed in the paper were able to obtain much information from the prescription. Our research has revealed preprocessing and text processing methods suitable for character recognition of Japanese prescriptions, as well as formats that are difficult to recognize, resulting in improvements in the efficiency of pharmacy business.