Research On Decision-making Suport For Urine Diagnosis In Tibetan Medicine

Tibetan medicine is an important part of Chinese ethnic medicine.It has a long history of more than 2000 years.At the same time,it is also a valuable diagnosis and treatment theory gradually recognized,developed and summarized by the ancient Tibetan people in the process of fighting against diseases.In thousands of years of practice,Tibetan medicine has summarized the theory of syndrome differentiation and disease identification with three means of observation,palpation and inquiry as the core,forming its own characteristics,and the most national characteristic of observation is Tibetan medicine urine diagnosis.Traditional Tibetan medicine urine diagnosis is observed and given diagnostic opinions by Tibetan doctors.Doctors’ diagnosis and treatment experience,professional skills and environmental factors will affect clinical diagnosis,resulting in strong subjectivity,low repeatability and lack of objective evaluation criteria.According to the clinical characteristics of Tibetan medicine urine diagnosis,this thesis studies the syndrome identification method for the information fusion of Tibetan medicine structured medical records,standardized urine diagnosis images and medical record images,and designs and implements the decision-making suport platform of Tibetan medicine urine diagnosis.The main contents are as follows:1.Aiming at the problem of syndrome identification for medical case of urine diagnosis in Tibetan medicine,the collection method of Tibetan medicine structured medical case is designed,and U-LIFT is proposed.Firstly,the medical case data is encoded,and the positive and negative sample sets are divided according to the basic syndrome types,which is transformed into a multi label classification problem.Then,the correlation structure information of the sample set is captured through unsupervised feature learning,and the specific features belonging to the syndrome labels are extracted for feature reconstruction.By training classifiers of each syndrome label,we completed the construction of multi label model for syndrome classification.The experimental results show that the accuracy of the Tibetan medical urine diagnosis data set is 81.71% and the accurate matching rate（accuracy of syndrome identification）is 64.21%.2.Aiming at the problem of syndrome identification for multi image fusion of urine diagnosis in Tibetan medicine,the collection scheme of urine diagnosis image are designed,and UIN is proposed.Firstly,the data set is newly expanded by using the data enhancement method,and then based on the pose normalized deep convolutional nets,according to the characteristics of urine diagnosis and data,the parallel convolution part and fusion convolution part of the network are introduced into SK convolution module and SE module respectively to realize the syndrome identification for standardized urine diagnosis images,The accuracy of the Tibetan medical urine diagnosis data is 67.35%.3.Aiming at the problem of syndrome identification for medical case and image fusion of urine diagnosis in Tibetan medicine,UFN is proposed.Combined with the structured medical case and standardized collected images in the urine diagnosis in Tibetan medicine,based on the convolutional neural network,the feature extraction part of U-LIFT and UIN is used to expand the basic network,the cross feature fusion based on attention mechanism is introduced,the text and image features are cross fused,and the syndrome identification results are output through the fully connected network.The recognition accuracy of Tibetan medical urine diagnosis data set was 70.80%4.Aiming at the problem that traditional Tibetan medical urine diagnosis is highly subjective and depends on doctors’ diagnosis and treatment experience,Tibetan medical urine diagnosis decision-making suport platform is designed and implemented based on Java and Java Script language.Vue,JS,React Native,Spring Boot and other frameworks are used to build web and mobile applications to make them more in line with the application scenario of intelligent decision-making function of Tibetan medical urine diagnosis.The platform has the functions of urine diagnosis database,urine diagnosis law analysis and urine diagnosis decision-making suport,which can assist Tibetan doctors to complete the diagnosis and treatment process more intelligently and accurately.