Skin Lesion Prediction for Skin Cancer Diagnosis Using ML Techniques

Abstract

Skin cancer (Malignant melanoma) is most rapidly growing cancer worldwide. Malignant melanoma is developed from pigment containing cells known as melanomas (melanocytes), normally appear on the skin but there is a very rare chance that it effects eyes and mouth. The diagnosis starts with medical screening, a biopsy and histopathological examination.

Clinically, it is very difficult to diagnose in early stages. Skin color is important feature for diagnosing malignant melanoma. Unfortunately, it is very difficult for physicians to diagnose skin cancer accurately and to differentiate between different skin cancer patients. Our research project focuses on early diagnosing of malignant melanoma. For this purpose, we will input the data set in different machine learning algorithms (Artificial Neural networks). Image processing and computer vision will be used for extracting comparable color features from selected skin lesion images. Best features will be analyzed and determined by doing complete statistical analysis which will help to classify skin lesions more accurately.