<title>Developing a predictive model of human skin coloring</title>

Abstract

A model of colour formation within human skin has been developed to aid the characterisation of pigmented skin lesions from their digitized colour images. The model is based on the Kublenka-Munk theory of scattering and absorption within inhomogeneous materials and the physics pertaining to their colour properties. By considering the skin to be a layered construction of such materials, the stratum corneum, epidermis, papillary dermis and reticular dermis, and by exploiting the physics related to the optical interface between these layers, the model generates all possible colours occurring within normal human skin. In particular, the model predicts that all skin colours have to lie on a simple curved surface patch within a three-dimensional colour space bounded by two physiologically meaningful axes, one corresponding to the amount of melanin within the epidermis and the other to the amount of blood within the dermis. These predictions were verified by comparing the CIE LMS coordinates of a representative, cross-racial sample of fifty skin images with the LMS coordinates predicted by the model. The results show that, within the predicted error bounds, the coordinates for normal skin colours do indeed lie on the curved surface generated by the model. Several possible applications of this representation are outlined, including images representing the melanin and blood components separately, as well as the possibility of measuring the Breslow thickness of melanocytic invasion within malignant melanoma.

From colour to tissue histology: Physics-based interpretation of images of pigmented skin lesions

“…Second, by generating images showing relative magnitudes of the histological entities, the lesion appearance can be correlated with its structure, thus providing an explanation as to why various skin diseases manifest themselves through particular visual signs. The model of normal skin colouration is representative of all the normal skins, irrespective of racial origin, age or gender [5]. The structure remains the same, and the only differences are in the magnitudes of the parameters c m , c h , and d pd .…”

“…We have implemented a two-flux model [5] based on Kubelka-Munk theory [6]. The model for the normal skin has three layers, corresponding to epidermis, upper papillary dermis (with prevalence of blood) and lower papillary dermis.…”

Spectral filter optimization for the recovery of parameters which describe human skin

“…The selection of optimal filters is a part of a broader method for the recovery of tissue parameters from tissue image data, originally developed by Cotton and Claridge [2], [17]. To provide an appropriate context, a brief description of the overall method is given.…”

“…D ESPITE considerable variability in skin hues across different human groups, the skin color depends primarily on three histological variables: The amount of melanin pigment, the amount of blood-borne hemoglobin pigments, and the thickness and structure of collagen in superficial skin layers [1], [2]. The idea of describing the entire range of skin coloration using only three parameters is very attractive.…”

“…Our earlier work [2], [5] developed a novel method for recovering histological skin parameters from image data using a physics-based model of image formation. An outline of that method is given in Section 3.…”