Melanoma Skin Cancer Classification Using Deep Learning Convolutional Neural Network

Abstract

In the recent years skin cancer skin cancer is emerging as one of the most complex diseases in which

diagnosis is very challenging. Melanoma is generally characterized by the uncontrolled growth of body

cells which might be caused due to prolonged exposure to UV rays produced by sun. Skin cancer can be

categorized as basal cell carcinoma, squamous cell carcinoma and melanoma among which melanoma is

considered as the most difficult to detect and if detected on time, melanoma is curable. Computer vision

and Image processing toolboxes plays a pivotal portion in the field of medical imaging and diagnosis and

is widely used. This paper focuses on a computer aided tool for skin cancer detection (i.e. melanoma).

Dermoscopic images are used as inputs to the CAD system which is subjected to further image processing

in which segmentation, feature extraction and classification is done to finally to differentiate between normal

and melanoma images.