This repository contains the code for ISIC 2018 Skin Lesion Classification Task:

https://challenge2018.isic-archive.com/task3/

Possible disease categories are:

Melanoma (MEL)
Melanocytic nevus (NV)
Basal cell carcinoma (BCC)
Actinic keratosis / Bowen’s disease (intraepithelial carcinoma) (AKIEC)
Benign keratosis (solar lentigo / seborrheic keratosis / lichen planus-like keratosis) (BKL)
Dermatofibroma (DF)
Vascular lesion (VASC)

The input images for training should in a folder containing class specific folder: MEL, NV, BCC, AKIEC, BKL,DF,VASC

python3 data_process.py will create the class specific folder structure, downlaod additional images from isic archive and merge them in a common class specific folder. The downloader is multi-threaded.

Data Collection: The task 3 has 10015 training images. We downloaded additional images from https://isic-archive.com/#images. we made a script that will