Design And Implementation Of Traditional Chinese And Western Medicine Diagnosis System Based On OCR

Dyslipidemia is a chronic disease that can cause diabetes,stroke and other cardiovascular diseases.At present,the problems of “difficult to see a doctor” and“expensive medical treatment” still exist,but the number of patients with hyperlipidemia is increasing.Therefore,the desires for timely diagnosis and treatment of patients is difficult to meet.Western medicine uses chemical drugs to treat hyperlipidemia,but it will produce side effects.With a long history and a great deal of theory,the treatment system of Chinese medicine for hyperlipidemia is constantly improving.At the same time,with the continuous development and continuous breakthroughs of technologies such as big data,deep learning algorithms,computer software and hardware,artificial intelligence and various industry fields continue to merge,providing new ideas for industry development.Based on the above situation,the diagnosis system of Chinese and Western medicine combines artificial intelligence technology with the medical field.By collecting the patient’s tongue characteristics,judging the patient’s degree of disease,and recognizing the patient’s inspection report picture,our system provide patients with online diagnosis and treatment services of hyperlipidemia.Our system uses the We Chat mini program as the mobile client,the web page uses the Vue.js framework.The back-end is divided into business processing server and inspection report processing server,using Spring Boot framework and Flask framework respectively.We use Mysql database to store data.The basic functional modules of this system include user center module,intelligent diagnosis module,user discovery module,chat and consultation module and system management module.I completed the design and implementation of the entire system,and realized the heartbeat mechanism and offline mechanism in instant chat through Web Socket in the chat and consulting module.The realization of the OCR technology in our system uses CTPN model to detect text boxes from inspection report table and CRNN model to recognize the text in text boxes from inspection report table.In order to improve the accuracy of detection and recognition of inspection reports,the algorithm innovations include:(1)Adding table extraction preprocessing and table restoration post-processing to the process of inspection report text detection.(2)Using Dense Net to extract features at the convolution layer and Using medical dictionary to correct the text at the transcription layer in the algorithm of inspection report text recognition.The m AP of inspection report text detection algorithm reached 92%,and the accuracy rate of inspection report text recognition algorithm reached 88%.Our system has been implemented and is in stable operation,which meets the needs of making it easier for people to get the diagnosis of hyperlipidemia with the combination of traditional Chinese and western medicine.Our system enhances the modern Chinese medicine service capabilities and provides more possibilities for the intelligent development of Chinese and Western medicine.