Classification of Melanoma and Nevus in Digital Images for Diagnosis of Skin Cancer

Abstract

Nevus and melanoma is recognized as deadly kind of skin disease. Sometimes it is tough to
tell apart, melanoma skin cancer is the most uncurable skin cancer compare to the other skin
cancers because it will spread easily to the other healthy part of the body, if it is not treated
well or not diagnosed earlier, it will easily spread to the other part of the body. The death rate
is higher compared to other cancers if it is detected sooner. The mortality rate becomes
higher, for diagnosing all patients the time and cost are high, so here we recommend the
image processing intellectual technique to differentiate and detect the infected disease from
nevus The initial step is removing the noise from the damaged skin from the attained image
using a Gaussian filter, SVM is used for the nevus and melanoma classification. Our main
goal is to check the efficiency of the suggested dissection technique, and obtain the features,
and evaluate the result with the other methods, In the current field we used. The suggested
technique is verified on the collected data by the overall figure of 402 collected data disease
image: 205 are nevus and melanoma are nevus 197. Our suggested technique achieved an
accuracy of 97%.