Integrated design of deep features fusion for localization and classification of skin cancer

• Published in: Pattern recognition letters Vol. 131; pp. 63 - 70
• Main Authors: Amin, Javeria, Sharif, Abida, Gul, Nadia, Anjum, Muhammad Almas, Nisar, Muhammad Wasif, Azam, Faisal, Bukhari, Syed Ahmad Chan
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.03.2020; Elsevier Science Ltd
• Subjects: Alexnet; Algorithms; Cancer; Cells; Classification; Datasets; Feature extraction; Localization; Melanoma; Principle component analysis (PCA); Skin cancer; Skin diseases; Skin lesions; VGG-16; Wavelet transforms; Alexnet; Principle component analysis (PCA); VGG-16; Cells; Melanoma
• ISSN: 0167-8655; 1872-7344
• DOI: 10.1016/j.patrec.2019.11.042

•Biorthogonal 2-D wavelet transform, and Otsu algorithm is used for segmentation.•The deep features of segmented images are extracted using Alex and VGG-16 model.•Extracted features are fused serially and selected optimal features using PCA.•The segmentation results are computed on ISIC 2018 datasets with ground truth.•Combined ISBI 2016, 2017, and PH2 dataset into single dataset for classification. The common fatal type of skin cancer is melanoma. Recently, numerous intelligent systems are used to detect skin cancer at an early stage. These systems are helpful for a dermatologist as a preliminary judgment to diagnose skin cancer. However, accurate skin lesion detection is an intricate task. This work comprises three main phases, firstly perform preprocessing to resize the images to 240 × 240 × 3 and convert RGB into L^* a^* b^* in which the luminance channel is selected. Secondly, Biorthogonal 2-D wavelet transform, Otsu algorithm are used to segment the skin lesion. Thirdly, deep features extracted from pre-trained Alex net and VGG16 and serially fused. The applied PCA for optimal features selection for classification into benign and malignant. The publically available datasets (PH2, ISBI 2016- 2017) are merged to form a single large dataset for the validated of proposed method. The results comparison is performed with the existing work which confirms that the proposed method classifies the skin lesion more accurately.