Advanced Processing Techniques for Detection and Classification of Skin Lesions

Abstract

The paper presents how the skin cancer in forms of melanoma can be identified based on the digital image processing of the Iesion. The solution is based on the extraction of seven features (deterministic and statistic type) from the image of a skin lesion: perimeter, area, diameter, fractal dimension, lacunarity, histogram of oriented gradients, and local binary patterns. Each feature has attached a specific classifier and the diagnosis is obtained by using a voting scheme in the final classifier. The experimental results on a free database demonstrate that the method provides a high accuracy.