L.C.P.S - Lung Classification and Pneumothorax Segmentation

The project was developed for the Deep Learning course with the dual objective of:

• classify rx images using four labels: covid, lung opacity, viral pneumonia and normal.
• identify the portion of the rx images containing the pneumothorax by image segmentation.

To achieve these goals, my team and I have implemented and trained various Deep Neural Networks:

• for the classification task: ResNet, VGG and Inception;
• instead, for the segmentation task: DeepLab and U-Net.