DermX experiments

Code for the experiments described in the DermX paper. Requirements must be installed with python 3.7.5.

Each of the two architecture folders, efficientnet and resnet contains the Jupyter notebooks used to run the hyper-parameter optimisation, the training of the five models considered in the paper, and the code needed to generate Grad-CAM visualisations. Grad-CAM notebooks use the methods stored in explanations/xgradcam.py.