Diagnosing Skin Cancer using Machine Learning Techniques

Abstract

Melanoma skin cancer is one of the deadly types of skin cancer and Biopsy is the method that is used to detect this cancer and success rate depends on performance of a trained doctor. The biopsy Process is very painful and requires considerable time. So, there is a need for a technique that can detect melanoma cancer that could avoid biopsy and that would be based on looking deep into skin cancer images. This paper has conducted a study on image classification of melanoma skin cancer using machine learning and various neural network techniques. The stages of the cancer image classification process melanoma skin in this study include the preprocessing process, segmentation, feature extraction with ABCD namely Asymmetry, Border Irregularity, Color Variation and Diameter. Subsequently, it's quantified that both the machine learning and neural network can be used for skin cancer diagnostics.