In this repository I train a deep nuerel network with a UNet architecture to detect up to 13 chest abnormalities from chest xrays.
The possible abnormalities are: Atelectasis, Calcification, Cardiomegaly, Consolidation, Diffuse Nodule, Effusion, Emphysema, Fibrosis, Fracture, Mass, Nodule, Pleural Thickening, Pneumothorax.
My final results (acheived with limited training xray images and GPU power) were:
Training and Validation Losses:
Accuracy on Validation Dataset:
