Embedded healthcare system based on bioimpedance analysis for identification and classification of skin diseases in Indian context

Abstract This chapter contributes toward real-time data collection, data management, system design, and analysis related to an embedded U-healthcare system for various skin diseases. The application domains of the developed system include diagnostic services, decision-support systems, and U-healthcare systems in real-world healthcare applications. The chapter explains a system developed for noninvasive identification and classification of skin diseases in the Indian context. The chapter describes the global scenario of skin diseases as well as in developing countries such as India. There are various challenges faced by dermatologists in the identification and classification of skin diseases. These are described with reference to state-of-the-art dermatological practices such as visual inspection, histopathological examination of a biopsy, and dermoscopy. The limitations of these subjective methods are discussed. The need to devise a noninvasive, reliable, simple, safe, and objective technique is described with reference to bioimpedance measurement. The chapter further describes a system developed for the measurement of human skin impedance. Skin impedance is measured with the help of a developed skin electrode. The requirements, specifications, and limitations of the developed skin electrode are described. The generation of a database of Indian skin diseases is described. The impedance of diseased and normal skin is measured with the help of the developed system and various impedance indices have been computed for an individual subject. The need of impedance indices for identification and classification of skin diseases is described. The identification of skin diseases requires discrimination between diseased and normal skin, which is explained with the help of the Wilcoxon signed rank test. The role of the Statistical Package for the Social Sciences (SPSS) is explained for computing the probability of similarity between diseased and normal skin. The possibility of the classification of skin diseases is described with the help of box and whisker plots and statistical measures of central tendency. The classification of skin diseases using a modular fuzzy hypersphere neural network has been explained. The performance of the proposed system for classification of skin diseases is explained with the help of confusion matrix and timing analysis.