CapsNet adaptation to melanoma detection.
To run this model you need:
1-A dataset with the features extracted by some other network (In this case VGG16) and save as .npy.
2-Run "create_name_list.py" with the directories of train, test and, validation, to create a list with all the names of the features files, and its respective labels.
3-Run "create_pickle_npy.py" to create batches of pickles to feed the Capsnet.
obs: You need to save as testX.npy...trainX.npy...validX.npy...
4-Run "capsulenet.py"
obs: The "ft.py" file do the extraction, u just need to load it with your onw keras model