Classification of Chest X-ray Images Using Deep Convolutional Neural Network.

In this study, we proposed an improved convolutional neural network (CNN) structure for the classification of normal images and pneumonia infection images in chest X-ray images. We choose Google's open source Inceptiont-Resnet-V2 network as the basic building block, connect it with the squeeze-and-excitation (SENet) module followed by a feature fusion layer. We use the opensource chest X-ray dataset on the kaggle platform to conduct experiments on the proposed framework. It is shown in the results that the proposed method can effectively improve the accuracy of chest X-ray image classification compared with the related CNN methods reported in the literature.