ABSTRACT
This chapter explores skin cancer classification methods using image processing and deep convolutional neural networks. Although instances of melanoma, a potentially fatal skin cancer, have been increasing at an alarming rate due to the high incidence of ultraviolet radiation, early diagnosis resulting in appropriate treatment measures has been found to improve survival chances. The naked eye is unable to assess the actual probability of a cancerous cell, but deep neural networks enable accurate identification of the presence of an underlying cancerous cell. We begin this chapter with training a deep neural network over thousands of images belonging to the classes benign and malignant. Interpreting the non-linear interactions in the LCNet model helps specify whether the image is benign or malignant. This AI-based automated clinical approach not only facilitates more efficient classification but is also tuned to avoid a large number of false positives and false negatives.
