In this paper, we introduce a technique of facial fattening (or slimming) simulation from a frontal facial image. In previous work, data-driven approaches enable a rapid design of fattening and slimming images that reflect realistic facial shapes; however, they focus only on facial surface such as facial landmarks, so it is difficult to consider any anatomical model such as skull structure. Then, we construct a novel facial database, which contains sets of frontal facial image and MRI image. Based on our database, we estimate 2D skull structure from the input image, and define a fattening (or slimming) rule and a skull-based constraint. As a result, we can generate a natural deformed image for fattening and slimming simulation.
|ジャーナル||Journal of the Institute of Image Electronics Engineers of Japan|
|出版ステータス||Published - 2017|
ASJC Scopus subject areas
- コンピュータ サイエンス（その他）