High Performance Method For Skin Roughness Detection Using Raman Spectroscopy

Background: the main goal of this study was to determine the features from the raman spectral data associated with skin roughness and to detect roughness from normal skin by using a proper classification method. material and methods: the raman spectral dataset was constructed from two classes of spectral data, 70 spectra of normal intact skin and 70 spectra of irritated rough skin. roughness irritation was induced by sodium dodecyl sulfate(sds) non-ionic surfactant applied daily on rat skin for a week. the spectra were obtained from upper legs and dorsal regions. some features related to specific bond vibrations of water, lipid and protein structures were extracted from the spectra. t-test statistical analysis was performed to determine whether the specified feature could discriminates two classes of spectral data. the reported efficient features from t-test analysis were applied to well-known classification methods such as linear discriminate analysis(lda) and k-nearest neighbors algorithm(knn for classification. classification performance was calculated using k-fold cross validation method for selecting the proper classifier and features. the statistical analysis of water content and lipid structures between two classes showed a significant difference by p-value < < 0.01, whereas alterations in features related to proteins were not remarkable between two classes of data. water content and lipid structures were the appropriate features for skin roughness detection. results and conclusion: the results from lda and knn for each extracted feature showed a maximum 80% accuracy and 76% specificity in classification of spectral data. in order to improve the efficiency of detection, combination of extracted features were applied to lda and knn classifier, which resulted in 85% accuracy and 87% sensitivity in classification.