ABCD Features Extraction-Based Melanoma Detection and Classification

Abstract

Malignant melanoma is said to be the most dangerous of all skin cancers. Around 232,000 were affected by it, and an approximate of 55,000 was recorded dead in year 2012. Although the cancer is classified in a lot of categories, but cancer generally is a representation of a cluster of diseases caused by the genetic disorder in cells of the body. The growth of defected cells is the major cause of cancer. These cells get affected because of differing causes and various degree of malignancy. Fortunately, if detected early, even malignant melanoma may be treated successfully. This paper also presented various methodologies. In research process, the number of melanoma clinical images has been considered, and it can be clearly seen that the Harris edge detection and preprocessing using median filtration and Otsu segmentation is observed to be the most preferred diagnosis. Feature extraction adds in it the method of symmetry detection, diameter detection, border detection and color to find variables of total dermoscopic value (TDS). Detection is done on the basis of range of this value. The experimental results show that the extracted features can be used to build a favorable classifier for melanoma detection.