Automated detection of melanocytes related pigmented skin lesions: A clinical framework. (May 2019)

• Record Type: Journal Article
• Authors: Pathan, Sameena; Gopalakrishna Prabhu, K.; Siddalingaswamy, P.C.
• Is Part Of: Biomedical signal processing and control. Volume 51(2019)
• Page Start: 59
• Page End: 72
• Publication Date: 2019-05
• Subjects: Benign lesions -- Dermoscopy -- Malignant lesions -- Melanocytic nevi -- Pigment network -- Shape -- Color -- Texture; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2019.02.013
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: An effective and automated system for melanoma diagnosis. Chroma based deformable models for localizing the skin lesions. The system takes into account the global and dermoscopic features for differentiating the pigmented skin lesions. Methodological classification between benign and malignant pigmented skin lesions using dynamic ensemble of classifiers. Abstract: A clinically oriented Computer-Aided Diagnostic (CAD) system is of prime importance for the diagnosis of melanoma, since the deadly disease is associated with high morbidity and mortality. Unfortunately, the development of CAD tools is hampered by several issues, such as (i) smooth boundaries between the lesion and the surrounding skin, (ii) subtlety of features between the melanoma and non-melanoma skin lesions, and (iii) lack of reproducibility of CAD systems due to complexity. The proposed system aims to address the aforementioned issues. First, the lesion regions are localized by incorporating chroma based deformable models. Second, the lesion patterns are analyzed to detect various dermoscopic criteria. Further, a robust ensemble architecture is developed using dynamic classifier selection techniques to detect malignancy. Quantitative analysis is performed on two diverse datasets (ISBI and PH2) achieving an accuracy of 88% and 97%, sensitivity of 95% and 97% and specificity of 82% and 100% for ISBI and PH2 datasets respectively.