Journal of Clinical & Experimental Dermatology Research
Open Access

Statistical analysis of skin texture and color for classification and disease diagnosis of human skin

Vitiligo and Skincare Physicians Meeting

September 15-16, 2016 Berlin, Germany

Gayatri Joshi

ACS College of Engineering, India

Posters & Accepted Abstracts: J Clin Exp Dermatol Res

Abstract :

Texture refers to visual patterns or spatial arrangement of pixels. Skin texture analysis plays a vital role in assessing the skin health and to diagnosis of skin disorders. Texture analysis is carried out by Structural, statistical, model based, transform based techniques. Statistical Texture analysis depend mainly on feature extraction which may be done using GLCM (Gray level Cooccurrence Matrix) and WDM (wavelength division multiplexing) techniques. The extracted features are used to classify texture. Skin texture can be analyzed using pixel intensity matrix parameters and GLCM. The research carried out by our team shows that Pixel intensity matrix often performs better than GLCM for analyzing skin texture. The following are the research outputs of our â�?�?Texture Analysis and disease diagnosisâ�? research team. The pilot study carried out by our texture analysis team of Bio-Medical dept. ACSCE, Bangalore. In my lecture I would flash the output of our research on skin texture analysis to classify the skin of different parts of the body. The pixel intensity matrix parameters perform better than GLCM in identifying the skin texture GLCM. Region of interest is then selected; the selected region of interest is an RGB image which is then converted to gray image. The ROI must have at least 400 to 600 pixels to get reliable results. For this work a set of normal mole and another set of abnormal moles (three images for each set) are taken for experimentation. The covariance, standard deviation, means, fv (maximum intensity pixel) are found for every image. From the values obtained a decision rule is framed to test the abnormality of the image of interest. There is a strong relation between the mole skin color and its pathological status i.e whether it is healthy or cancerous. When the mole is cancerous one its color highly differs from that of the one which is healthy that is why in the golden rule (A, B, C, D, and E) parameter diagnosis method used in skin cancer diagnosis the color plays a vital role.

Biography :

Email: gayitrijoshi@gmail.com