The spatial distributions of melanin and hemoglobin in human skin are separated by independent-component analysis of a skin color image. The analysis is based on the skin color model with three assumptions: (1) Spatial variation of color in the skin is caused by two pigments, melanin and hemoglobin; (2) the quantities of the two pigments are mutually independent spatially; and (3) linearity holds among the quantities and the observed color signals in the optical density domain. The results of the separation agree well with physiological knowledge. The separated components are synthesized to simulate the various facial color images by changing the quantities of the two separated pigments.

/pdf/independent-component-analysis-of-skin-color-image-2fjk6eaot6.pdf

Independent-component analysis of skin color image

A color image processing apparatus reproduces color data of a subject at higher accuracy by estimating the spectral reflectance of the subject from the color chip basis function vn (λ) which can expand the spectral reflectances of all color chips of the reference chart and the expansion coefficients B of the color chips The color image processing apparatus uses an image processing device to which a subject shooting signal Q shot by the color image input device and a reference chart shooting signal G taken from a plurality of color chips each having a spectral reflectance, is known

Color image processing apparatus

Most face databases are meant for pattern recognition use. We report the creation of a unique face database which may be used for color-related studies of faces. We describe it as physics- based because it not only contains color images but also data on color image formation such as spectral reflectance measurements from facial skin, illuminant spectral power distribution, and camera spectral response. We demonstrate the usefulness of the database by an approach we developed to color correct face images taken under different illumination conditions. © 2000 SPIE and IS&T.

Physics-based face database for color research

Quantitative analysis of human skin color is needed in the medical and cosmetic fields. Because of the strong light scattering by biological tissues, however, analysis of skin color has not yet been fully successful. Human skin color is dominated by the colors of blood and melanin which are modified by scattering. Exposure of human skin to hot water or UV-ray changes its color because the absorbance spectrum of reflection from the human skin, and the changes in the absorbance spectra of reflection reflect the absorbance spectra of blood and melanin which are the absorption spectrum distorted by scattering. By applying the modified Lambert-Beer law, the absorbance spectrum of reflection from human skin can be expressed proportional to those distorted absorbance spectra of blood and melanin. Multiple linear regression analysis is successfully used to reproduce the absorbance spectrum of reflection from human skin from the distorted absorbance spectra of blood and melanin.

Explanation of Human Skin Color by Multiple Linear Regression Analysis Based on the Modified Lambert-Beer Law

Measurement of melanin and blood concentration in human skin is needed in the medical and the cosmetic fields because human skin colour is mainly determined by the colours of melanin and blood. It is difficult to measure these concentrations in human skin because skin has a multi-layered structure and scatters light strongly throughout the visible spectrum. The Monte Carlo simulation currently used for the analysis of skin colour requires long calculation times and knowledge of the specific optical properties of each skin layer. A regression analysis based on the modified Beer-Lambert law is presented as a method of measuring melanin and blood concentration in human skin in a shorter period of time and with fewer calculations. The accuracy of this method is assessed using Monte Carlo simulations.

https://iopscience.iop.org/article/10.1088/0031-9155/46/9/309/pdf

Melanin and blood concentration in a human skin model studied by multiple regression analysis: assessment by Monte Carlo simulation

We propose a new method to predict the images on a CRT and a hardcopy of skin color under various illuminants. Spectral reflectance of human skin is analyzed by principal component analysis and it is shown that the spectral reflectance can be estimated by three basis functions. The estimation allows colorimetric color reproduction without colorimetric measurements for each illuminant. The proposed method is verified by several skin color patches taken by a calibrated HDTV camera. The method is also applied to practical facial pattern images.

Principal component analysis of skin color and its application to colorimetric color reproduction on CRT display and hardcopy

In this article, we study a color reproduction method to match the appearance of skin color images on a CRT with a hardcopy under various viewing illuminants. Three color appearance models, von Kries, Fairchild, and CIELAB, were applied to predict color reproduction on a CRT display. The optimum color appearance model to predict color reproduction was estimated by psychophysical experiments based on memory matching. As a result of the experiment, differences in visual perception between skin color patches and facial pattern images are also described.

Prediction of color reproduction for skin color under different illuminants based on color appearance models

We improved the incomplete chromatic adaptation model proposed by Fairchild to reproduce facial pattern images under various illuminants. We introduced weighting coefficients in the modified fundamental chromaticities of the adapting stimulus used in the Fairchild model to improve the color reproduction. Psychophysical experiments using memory matching technique were performed to select the optimum weighting coefficients of adapting stimulus for facial pattern images. The improved model with the optimum coefficients was compared with other color appearance models, von Kries, RLAB, LLAB, and colorimetric color reproduction under three illuminants; illuminant A (2837 K), Daylight (6047 K) and Cool White (3957 K). As a result, it was shown that the improved model performed significantly well for facial pattern images if compared with other models.

F. Hideki Imai

Papers

Principal component analysis of skin color and its application to colorimetric color reproduction on CRT display and hardcopy

Prediction of color reproduction for skin color under different illuminants based on color appearance models

Improvement of incomplete chromatic adaptation model for facial pattern images