Chest X-ray Pneumonia Detection Based on Convolutional Neural Networks
Zebin Jiang
Abstract
For non-professionals, it is difficult to tell whether the patient has pneumonia through chest X-ray images. If the convolutional neural network is used to handle this task, it will improve the diagnosis efficiency of pneumonia and reduce the workload of doctors. Therefore, this research is of great significance. Multiple variants of convolutional neural networks are used to handle chest X-rays pneumonia detection tasks, namely InceptionResNetV2, Xception, DenseNet201, and VGG19. Besides, the whole process from dataset selection, dataset processing to chest X-rays pneumonia detection is described. Experimental results show that chest X-rays pneumonia detection based on convolutional neural networks can improve training speed and detection accuracy, and the highest accuracy reaches 94.20%.