Detection of blue-whitish veil in melanoma using color descriptors

Recent studies have shown a dramatic increase in the incidence of malignant melanoma. Thus, it is essential to educate the public about risk factors and augment screening for melanoma to reduce fatalities. Common diagnostic criteria used by dermatologists specialized in dermoscopy are based on the detection of certain texture and color features of a lesion. In this paper, we present a novel algorithm for detecting the socalled blue-whitish veil color feature. Our results show that when the blue-whitish veil is combined with the regression feature can provide sensitivity and specificity of 95.64% and 72.30%, respectively. Therefore detecting blue-whitish veil accurately is of paramount importance in systems for automated detection of melanoma.