Research And Development Of Auxiliary Diagnosis System For Chest Diseases Based On Deep Learning

As of 2019,there are 2749 tertiary hospitals in China(including 1516 tertiary hospitals).There are different number of tertiary hospitals serving local people in different provinces and cities,but there are still two problems: first,the number of uneven distribution.The number of tertiary hospitals in developed areas is more than that in underdeveloped areas;Second,the level gap is too large.Although they are all tertiary hospitals,there is a big gap in the level of hospitals in different regions.According to statistics,there are 34 provincial administrative regions in China,among which Tibet,Anhui,Inner Mongolia,Hebei and Gansu are the top five provinces in terms of the proportion of patients lost;Shanghai,Beijing,Jiangsu,Zhejiang and Guangdong were the top five provinces in the proportion of patients inflow.It can be seen that the five provinces with the highest patient loss rate are basically concentrated in the underdeveloped areas in the central and western regions,while the provinces with the highest patient inflow rate are mainly concentrated in the eastern developed areas.To a certain extent,people in less developed areas rarely have the opportunity to communicate with high-level doctors in developed areas.People need to bear more transportation costs and time costs to access high-quality medical resources.In addition,when people go to the hospital for treatment,they may have a long queue time.When they get the examination results,they may face the situation that the doctor is off duty.As a result,the obtained examination results cannot be diagnosed and they need to go to the hospital another day,which also increases the transportation and time costs of people.With the continuous improvement of the level of social information,through artificial intelligence and network,people can more easily access to high-quality medical resources,more quickly and conveniently get the diagnosis results has gradually become a possibility.Based on the above background,this paper designs a system construction plan that takes online light diagnosis of chest disease X-rays as the core function,and integrates modules such as online light consultation,physical examination center,and healthy community.On the mobile terminal,the corresponding We Chat applet was designed and implemented,achieving the purpose of using this system in any environment and location.The main tasks completed in this paper are as follows:1.In order to solve the problem that mobile phones and other handheld devices are affected by the light,angle of the shooting environment and the performance of the device itself,which will affect the accuracy of lesion recognition after the image is uploaded to the system,a random chexnet model is proposed.After the training samples are input into the network,through the data enhancement method of angle rotation,adding shadow,blur and brightness to the samples,the training samples are added,and then the samples are input into the deep convolution neural network,and finally the thermal map is output.The experimental results show that the average recognition rate of random chexnet is close to that of chexnet in a few sample data sets,and the recognition ability of random chexnet is better than that of chexnet in the images with shadow,blur and brightness.2.Designed an architecture scheme of an auxiliary diagnosis system for chest diseases based on deep learning.The system adopts the micro-service architecture for design and implementation.The system consists of user front-end website,back-end management system and other content.The front-end website includes modules such as homepage display,physical examination center,light X-ray diagnosis,online light consultation,and healthy community;the back-end management system includes modules such as X-ray management,physical examination center management,healthy community management,and user management.3.In response to the ever-increasing use of We Chat mini-programs,we have designed and implemented "smart and healthy" mini-programs on the mobile terminal.This small program provides the same functions as the PC terminal,such as X-ray light diagnosis,online light consultation,healthy community,and physical examination center.It achieves the purpose of allowing users to use the system in any environment.