| Title |
Face Age Simulation based on Pyramid-cGAN and Aging Function |
| Authors |
최성은(Sung Eun Choi) ; 홍유진(Yu-Jin Hong) ; 김익재(Ig-Jae Kim) |
| DOI |
https://doi.org/10.5573/ieie.2021.58.8.65 |
| Keywords |
Face age simulation; Face age synthesis; GAN; Aging function |
| Abstract |
Recently, research on facial age simulation based on Generative Adversarial Networks (GAN) is actively being conducted. Although facial texture aging is successfully represented by the GAN-based facial age simulation method, there is a problem that the facial shape aging cannot be sufficiently represented. Therefore, we propose an age simulation method that combines GAN and aging function to sufficiently represent the age of facial texture and shape. In the proposed method, the facial texture is transformed based on Pyramid-cGAN, which combines Pyramid-architecture discriminator, age loss, and identity-preserved loss with Conditional-GAN (cGAN). The facial shape is transformed using the PCA features of facial landmarks and aging function. The final age simulation face is generated by image warping of texture and shape which are transformed by Pyramid-cGAN and aging function respectively. The experimental results show that the proposed method simulates a face more accurately and realistically than previous methods. |