Fusion of structural and textural features for melanoma recognition

• Published in: IET computer vision Vol. 12; no. 2; pp. 185 - 195
• Main Authors: Adjed, Faouzi, Safdar Gardezi, Syed Jamal, Ababsa, Fakhreddine, Faye, Ibrahima, Chandra Dass, Sarat
• Format: Journal Article
• Language: English
• Published: The Institution of Engineering and Technology 01.03.2018; IET; Wiley
• Subjects: cancer; cancers; dermoscopy; Engineering Sciences; feature extraction; image fusion; image recognition; local binary pattern operator; medical image processing; melanoma recognition; Review Article; Signal and Image processing; structural features; support vector machine; support vector machines; SVM classifier; textural features; wavelet transforms; support vector machines; wavelet transforms; image fusion; structural features; local binary pattern operator; textural features; support vector machine; feature extraction; cancers; cancer; melanoma recognition; SVM classifier; medical image processing; image recognition; dermoscopy
• ISSN: 1751-9632; 1751-9640; 1751-9640
• DOI: 10.1049/iet-cvi.2017.0193

Melanoma is one the most increasing cancers since past decades. For accurate detection and classification, discriminative features are required to distinguish between benign and malignant cases. In this study, the authors introduce a fusion of structural and textural features from two descriptors. The structural features are extracted from wavelet and curvelet transforms, whereas the textural features are extracted from different variants of local binary pattern operator. The proposed method is implemented on 200 images from ${\rm P}{\rm H}^2$PH2 dermoscopy database including 160 non-melanoma and 40 melanoma images, where a rigorous statistical analysis for the database is performed. Using support vector machine (SVM) classifier with random sampling cross-validation method between the three cases of skin lesions given in the database, the validated results showed a very encouraging performance with a sensitivity of 78.93%, a specificity of 93.25% and an accuracy of 86.07%. The proposed approach outperforms the existing methods on the ${\rm P}{\rm H}^2$PH2 database.