Analysis of skin surface roughness by visual assessment and surface measurement

Abstract

In this study, we analyze the surface roughness of human skin using human visual assessment and statistical features of three-dimensional shape data and goniometric reflectance data. We use eight skin replicas taken from the cheeks of women. First, we perform human visual assessment to obtain the roughness rankings. Second, we measure three roughness parameters by currently used methods. These parameters show no significant correlation with the roughness rankings. Third, we measure the skin surface shape to calculate surface normal vectors. We show that the surface normal distribution of the skin is isotropic and Gaussian, and the standard deviation has a good correlation with the roughness rankings. Finally, we analyze the goniometric reflectance data to approximate the surface reflection using the Oren-Nayar model. We find that the standard deviation estimated using this model corresponds to the roughness rankings. Thus, this parameter can be effectively used for describing skin roughness.