一种基于深度学习的移动端胆石病智能诊断系统

An Intelligent Diagnostic System for Cholelithiasis Deployed on Mobile Terminal and Based on Deep Learning

传统医疗方式大都将采集到的数据发送到本地或远程服务器,然后将处理的结果发送给用户。论文设计并研发了一种基于深度学习的移动端胆石病智能诊断系统,可以离线部署在具有安卓系统的移动端设备,并在本地进行胆石病医疗图像识别工作。该系统采用直方图均衡化进行图像预处理以及卷积神经网络进行病灶特征的提取。通过联合编译Python和Java代码来调用模型在移动终端的应用。实验表明,该系统能够在确保准确率高达94.8%的前提下快速完成胆石病识别过程。

Traditional medical methods usually send the collected data that we to local or long-range servers and then send the results to users.This paper designs and implements an intelligent diagnostic system for cholelithiasis based on the Android platform.The system can be deployed offline on a mobile device equipped with an Android system,and performs the work of medical image recognition on the local mobile terminal.This paper uses histogram equalization for preprocessing the image and lightweight convolutional neural network for extracting feature of image and recognizing cholelithiasis.This paper compiles Java and Python to adapt to the application of model on the mobile terminal.Experiments show that the system can quickly complete the recognition process of cholelithiasis on the premise of ensuring the accuracy rate of up to 94.8%.