Automatic Skin Lesion Segmentation using SegNet, a Deep Learning architecture with certain extra requirements. Keeping the pre- and post-processing of the photos to a minimum is the secondary goal. The dermoscopic pictures in the PH2 dataset, which are part of the limited amount of images used to train the proposed model, were manually segmented.
Dataset
Model Architecture
