This project explores diagnosing pneumonia through chest x-ray images with the help of deep learning (TensorFlow & Keras) and various data augmentation methods. The dataset used for this project is part of a Kaggle challenge and can be found here.
The neural network used is written in Keras and consists of three convolution blocks as well as a max pool layer for each one of them. On the top there's a fully connected layer with 512 units which is activated using relu-activation. We also added Dropout layers to two of the convolution layers to reduce over fitting due to the scarcity of data.
The images in the dataset were scaled to 150x150 pixels and converted from RGB to grayscale before they were fed to the model.
The data augmentation methods that were tested were Cutmix, Mixup and Cutout, as well as combinations of these.
