Yet another PyTorch implementation of the CheXNet algorithm for pathology detection in frontal chest X-ray images. This implementation is based on approach presented here. Ten-crops technique is used to transform images at the testing stage to get better accuracy.

The highest accuracy evaluated with AUROC was 0.8508 (see the model m-25012018-123527 in the models directory). The same training (70%), validation (10%) and testing (20%) datasets were used as in this implementation.

• Python 3.5.2
• Pytorch
• OpenCV (for generating CAMs)

• Download the ChestX-ray14 database from here
• Unpack archives in separate directories (e.g. images_001.tar.gz into images_001)
• Run python Main.py to run test using the pre-trained model (m-25012018-123527)
• Use the runTrain() function in the Main.py to train a model from scratch

This implementation allows to conduct experiments with 3 different densenet architectures: densenet-121, densenet-169 and densenet-201.

• To generate CAM of a test file run script HeatmapGenerator

The highest accuracy 0.8508 was achieved by the model m-25012018-123527 (see the models directory).

The training was done using single Tesla P100 GPU and took approximately 22h.