Image Classification for Pneumonia and the Normal Based on Convolutional Neural Network

Our model focused on distinguishing pneumonia X-ray images from normal ones. Based on VGGNet16, some layers were reduced in our model; meanwhile, the dropout function and batch normalization were added to avoid the overfitting. Adam was used as the optimizer, and Binary Cross Entropy was used as the loss function. Aiming to evaluate the model, we used different metrics based on the confusion matrix to analyze its performance, including the recall of model, the precision and the F1 score. Our model’s final accuracy reached 86.54%.