Automated skin lesion analysis based on color and shape geometry feature set for melanoma early detection and prevention

Abstract

Melanoma incidence rates have been increasing for the past three decades. Most people diagnosed with non-melanoma skin cancer have higher chances to cure, but melanoma survival rates are low comparing to other skin cancer types. It is important to note that one in five Americans will develop skin cancer in their lifetime, and on average, one American dies from skin cancer every hour. A system to prevent this type of skin cancer is being awaited and is highly in-demand. Early detection of melanoma is one of the major factors to increase the chance of cure significantly. Malignant melanomas are asymmetrical and have irregular borders with rages and notched edges, so analyzing the shape of the skin lesion is important for melanoma early detection and prevention. In this paper, we introduce an automated skin lesion segmentation and analysis for early detection and prevention based on color and shape geometry. The system further incorporates other feature sets such as color to determine the lesion type. In our proposed system, we used PH2 Dermoscopy image database from Pedro Hispano Hospital for the development of our system and for testing purposes. This image database contains a total of 200 dermoscopy images of lesions, including normal, atypical, and melanoma cases. Our approach of analyzing the shape geometry and the color will be helpful to detect atypical lesions before it grows and becomes a melanoma case.