基于卷积神经网络的食管癌图像智能诊断系统设计

Research and Design of Image Intelligent Diagnosis System for Esophageal Cancer Based on Convolutional Neural Network

目的:开发食管癌智能诊断系统,提升食管癌图像诊断的识别率和医生工作效率,同时提升基层医疗机构的食管癌图像诊断水平。方法:联合腾讯公司,采集多年来食管癌相关医学图像,通过数据标注、图像预处理、数据增强、深度学习等步骤,建立基于积神经网络的食管癌图像智能诊断系统,辅助医生进行智能诊断。结果:基于卷积神经网络的食管癌图像智能诊断系统在医院成功应用,得到一线医生的广泛好评。结论:此系统有利于基层医生全面提高食管癌诊断准确率,降低食管癌患者死亡风险。

Objective： Developing an intelligent diagnosis system for esophageal cancer, improve the recognition rate of image diagnosis of esophageal cancer and the efficiency of doctors, and at the same time improve the diagnostic level of esophageal cancer in primary medical institutions. Methods： Collecting medical images of esophageal cancer, after collect medical images related to esophageal cancer for many years, through data annotation, image preprocessing, data enhancement, deep learning and other steps, finally established image intelligent diagnosis system for esophageal cancer. Results： The image intelligence diagnosis system of esophageal cancer based on convolutional neural network was successfully applied in hospital and received wide acclaim from first-line doctors. Conclusion： The use of this system will help grass-roots doctors to improve the diagn^ostic accuracy of esophageal cancer and reduce the risk of death from esophageal cancer.