A CNN model to distinguish whether a skin lesion is malignant or benign: cancerous or not. Trained on 1500 images for both benign and malignant classes attained from the ISIC skin condition dataset with over 84% accuracy and 93 F1 score.

The purpose of this project is to correctly classify the skin lesion category present in an image, by building the most accurate ML model. 4 different CNN architectures were used and evaluated.