%0 Conference Proceedings
%T Improving the attractiveness of faces in images
%D 2015
%A Leite, Tatiane Silvia,
%A Martino, Josť Mario De,
%@affiliation State University of Campinas
%@affiliation State University of Campinas
%E Papa, Jo„o Paulo,
%E Sander, Pedro Vieira,
%E Marroquim, Ricardo Guerra,
%E Farrell, Ryan,
%B Conference on Graphics, Patterns and Images, 28 (SIBGRAPI)
%C Salvador
%8 Aug. 26-29, 2015
%I IEEE Computer Society
%J Los Alamitos
%S Proceedings
%K facial attractiveness, facial geometry, skin texture, machine learning.
%X Advertising images increasingly require attractive faces to attract the publics attention. Several studies have been conducted to enhance facial attractiveness in images. While some researchers suggest changes in geometrical shape, others advocate modifying the appearance of the facial skin; however, there have been few attempts to explore the possibility of combining both techniques. This paper sets out a novel method of doing this: facial geometry and skin texture modifications. Our method, which is based on supervised machine learning techniques, is able to improve the attractiveness of faces in images while preserving the original features of the picture. We also demonstrate the effectiveness of this combination by carrying out two different evaluations. Accordingly, we analyze the significance of each change that is designed to improve attractiveness by comparing the original image with a) the image in which only the facial geometry has been modified, b) the image in which only the texture skin has been modified and finally c) the image with both modifications. Our results reveal that the combination of geometric and skin texture modifications results in the most significant enhancement. It also demonstrates that modifications to the skin texture can be regarded as more important to obtain an attractive face than changes to the facial geometry. Additionally, evaluations are provided to quantify the gain in facial attractiveness and it should be pointed out that our method is the first to employ these, since there are no references to such tests in the literature.
%@language en
%3 Improving_the_attractiveness_of_faces_in_images.pdf