A Two-Picture Acquisition System for 3D Geometry and Light Reflection Modeling of Faces for Cosmetic Applications
Hussam Yousef, Céline Loscos, Régis Huez, Michel Herbin · 2013
Objectifying the visual effects of cosmetic products is a major concern for both manufacturers and consumers.One of the most complex effects of cosmetic products is the transparent aspect of the top layers of the skin, called translucency.Most evaluation methods of this phenomenon are still very subjective due to the complex visual quantification.For evaluating the effects of cosmetic products on skin translucency, we have developed an acquisition system to recover facial geometry and translucency coefficients.The equipment is designed to be low-cost and portable, and our reconstruction procedure needs only two photographs.The algorithm for 3D reconstruction uses one photograph of the face of a human subject on which was projected colored structured light to extract a 3D surface.An image-based approach deduces an RGB translucency value for a skin patch from a unique non-calibrated photograph.The proposed system was proven robust by means of experimentations conducted on 20 subjects and three different cosmetic products.We show that an image-based evaluation procedure is sufficient to consistently differentiate translucent effects of a range of cosmetics on a same skin, even from a unique non-calibrated image.The extracted parameters can be used for assessment of the effect of a product or for simulating this effect on a face.