A Study on RGB Color Extraction of Psoriasis Lesion using Principle Component Analysis (PCA)

This paper presents the application of Principle Component Analysis (PCA) on RGB color feature extraction of psoriasis image lesions. PCA is used to reduce the dimensionality of data for image compression and is commonly used dataset in micro array research as a cluster analysis tool. In this work, several clinical psoriasis lesion groups are been studied for digital RGB color features extraction where five sets of skin lesion images are digitally captured under standard and control environment. The identified regions of interest (ROI) of these lesions' images are processed to quantify the conventional normal and differential reflectance indices in RGB color model. Statistical performances of the RGB model before and after applying PCA are analyzed and compared with respect to two samples sizes; the original 90 samples and the size reduction of 73 samples after PCA was applied. Result performances are concluded by observing the error plots with 95% confidence interval and outcome of the inference statistical t-tests. The tests outcomes have shown that B component for differential method can be used to distinctively classify erythroderma, guttate and plaque with p-value reaching more than 0.8.