Micro-Geometric Skin Simulation for Face Impression Analysis

Abstract

Condition of skin is an important factor for the impression of faces. Well-cared skins give shiny and beautiful impression, while dry and rough skins loss the gloss and therefore they are less attractive. These differences often attribute to the difference of their micro-geometry. Such micro-geometry of human skins consists of furrows forming a mesh, ridges surrounded by the furrows, and pores. This paper presents a technique for micro-geometry simulation of human skins. The technique first generates pores forming a pattern of a well-aligned triangular grid. It then divides the skin region by applying the Delaunay triangulation algorithm to connect the pores. It treats the triangles as ridges, and the edges as furrows. Finally, the technique divides the pattern into a finer triangular mesh, so that it can finely represent 3D geometry of pores, ridges, and furrows. This paper also presents a technique for image recognition of micro-geometry parameters of real human skins. It features two-steps of template-matching- like algorithms: the former one acquires sizes of pores, and the latter one acquires widths and directionality of furrows. These results are consumed by the skin simulation technique as input parameters of the micro-geometry.