Estimation of Melanin Concentration, Blood Concentration, and Oxygen Saturation in Skin Tissue Layers with Different Depths Unaffected by Shading

摘要

Measurement of melanin concentration, blood concentration and oxygen saturation in skin tissue is highly importanttechnology in medical and cosmetic fields. Most previous studies neglect depth inhomogeneity of blood concentrationdistribution in skin tissue. However, skin tissue has a complex layered structure, and it is known that distribution of bloodconcentration differs greatly depending on the depth. Although a method has been proposed considering the inhomogeneityof blood distribution in depth direction, it is impossible to measure oxygen saturation. In addition, there is a limitation thatshading influences the estimation as noise in any of the methods. Therefore, in this study, we propose a method to measureskin components concentration, which are melanin concentration, blood concentration, and oxygen saturation, in layerswith different depths of skin tissue without influence of shading. First, to construct a dataset of absorbance correspondingto the skin components concentration we construct an optical multi-layered model of skin tissue and simulate lightpropagation and attenuation by light scattering and absorption with Monte Carlo simulation method. Each absorbancevalue is converted to a relative value based on an averaged absorbance value of all wavelengths to remove influence ofshading. Next, we create a model describing an absorbance value corresponding to the skin components concentrationwith high dimensional multivariate regression based on the created dataset. This model enables to estimate the skincomponents concentration from spectral information without influence of shading. The proposed method is evaluated bysimulation and shown to have state-of-the-art performance.