Analysis And Research The Lung X-ray Images Based On Big Data

At present,the most important X-ray film review relies on manual review,but the growth rate of radiological image is much higher than radiologists,in China,the growth rate of radiological image is 30%,the growth rate of the radiologists is only 4.1%.The unbalance growth between radiological images and radiologists has made the radiologists face too many pressures every day,so that the misdiagnosis may increase.Meanwhile,different radiologist may give different examination results to same X-ray image.For those reasons,scientific research and medical institutions hope to use big data and artificial intelligence technology to assist radiologists in checking different medical images to improve the accuracy of medical image recognition and reduce the pressures of the radiologists,achieving the purpose of early detection and early treatment.In this project,using big data technology to store lung X-ray image data by distributed storage way on cloud platform,and using deep learning technology to extract feature from a large number of lung X-ray images automatically.Compared the accuracy of lung X-ray detection by AlexNet,MobileNets and VGG16 neural network model,choosing the better one as the basis of algorithm improvement for twice optimization named VGG-X and VGG-X1.VGG-X network model replaces the standard convolution into depthwise separable convolution on the basis of VGG16.VGG-X1 network model adds BN layer after pooling layer of VGG-X.Recognizing the consolidation,pulmonary edema,effusion,emphysema,fibrosis,hernia,infiltration,mass,pneumonia and normal symptoms by VGG-X1.The experimental results show that the VGG-X1 neural network model plays better in recognizing the lung X-ray images,it can achieve an average accuracy more than 85%.Compared with VGG16,VGG-X1 has higher accuracy,needs fewer parameters and less time.What is more,the scalability and security of data storage is improved by big data technology.