Diagnostic techniques for improved segmentation, feature extraction, and classifi cation of malignant melanoma

• Published in: Biomedical engineering letters pp. 171 - 179
• Main Authors: Hyunju Lee, Kiwoon Kwon
• Format: Journal Article
• Language: English
• Published: 대한의용생체공학회 01.02.2020
• Subjects: 의공학
• ISSN: 2093-9868; 2093-985X
• DOI: 10.1007/s13534-019-00142-8

A typical diagnosis of malignant melanoma involves three major steps: segmentation of a lesion from the input color image,feature extraction from the separated lesion, and classifi cation to distinguish malignant from benign melanomas based onfeatures obtained. We suggest new methods for segmentation, feature extraction, and classifi cation compared. We replacededge-imfi ll method with U-Otsu method for segmentation, the previous features with new features for the criteria ABCD(asymmetry, border irregularity, color variegation, diameter) criteria, and the median thresholding with weighted receiveroperating characteristic thresholding for classifi cation. We used 88 melanoma images and expert’s segmentation. All thethree steps in the suggested method were compared with the steps in the previous method, with respect to sensitivity, specificity, and accuracy of the 88 samples. For segmentation, the previous and the suggested segmentations were also comparedassuming the skin cancer expert’s segmentation as a ground truth. All three steps resulted in remarkable improvement inthe suggested method. KCI Citation Count: 0