Computational methods for pigmented skin lesion classification in images: review and future trends

Abstract

Skin cancer is considered as one of the most common types of cancer in several countries, and its incidence rate has increased in recent years. Melanoma cases have caused an increasing number of deaths worldwide, since this type of skin cancer is the most aggressive compared to other types. Computational methods have been developed to assist dermatologists in early diagnosis of skin cancer. An overview of the main and current computational methods that have been proposed for pattern analysis and pigmented skin lesion classification is addressed in this review. In addition, a discussion about the application of such methods, as well as future trends, is also provided. Several methods for feature extraction from both macroscopic and dermoscopic images and models for feature selection are introduced and discussed. Furthermore, classification algorithms and evaluation procedures are described, and performance results for lesion classification and pattern analysis are given.

Skin lesion computational diagnosis of dermoscopic images: Ensemble models based on input feature manipulation

“…An overview of computational methods for pigmented skin lesion classification in images, which addresses the feature extraction and selection, and classification steps, is presented in Oliveira, et al [12]. The ensemble of classifiers based on input data manipulation has been recently adopted for skin lesion classification to achieve better results than single classifiers.…”

“…In this study, six feature selection algorithms were applied to generate different feature subsets for the ensemble of classifiers; namely, Pearson's correlation coefficient [35], gain ratio-based feature selection (GRFS) [5], information gain-based feature selection [35], relief-F [36], principal-component analysis (PCA) [37] and correlation-based feature selection (CFS) [38]. These algorithms have been commonly used for skin lesion feature selections [12] since they have several advantages, such as computationally efficiency, are simple and fast algorithms, independent evaluation criteria, and have the ability to overcome over-fitting.…”

Search for Resolution Invariant Wavelet Features of Melanoma Learned by a Limited ANN Classifier

“…Classifier performance results from other existing melanoma CAD systems can be found in [55]. The last review [3] gives a vast look at different feature types and classification methods. It reports and classifies the 2007-2015 studies paying attention at refering them to the global and local patterns.…”

“…Early diagnosis of this tumor is a lifesaving factor. Lack of specialists, too late detections and an increased melanoma morbidity rate have become a medical problem for some time and, on that account, a challange for computer assisted diagnosis (CAD) [1][2][3]. The standard therapy, biopsy, is not feasible for all the patients due to treatment costs and some health reasons.…”

Pattern Recognition in Macroscopic and Dermoscopic Images for Skin Lesion Diagnosis

“…Furthermore, 9,730 deaths from melanoma were estimated for the same year [1]. Computational systems have been proposed in order to assist dermatologists in skin cancer diagnosis, or even to monitor skin lesions [2,3]. Image acquisition, pre-processing, segmentation, feature extraction and selection, and classification are fundamental steps commonly found in computational systems for diagnosing skin lesions.…”