Section: New Results

From Attribute-labels to Faces: Face Generation using a Conditional Generative Adversarial Network [53], [54]

Participants : Yaohui Wang, Antitza Dantcheva, Francois Brémond.

Keywords: Generative Adversarial Networks, Face generation

Facial attributes are instrumental in semantically characterizing faces. Automated classification of such attributes (i.e., age, gender, ethnicity) has been a well studied topic. We here seek to explore the inverse problem, namely given attribute-labels the generation of attribute-associated faces. The interest in this topic is fueled by related applications in law enforcement and entertainment. In this work, we propose two models for attribute-label based facial image and video generation incorporating 2D (see Figure 10) and 3D (see Figure 11) deep conditional generative adversarial networks (DCGAN). The attribute-labels serve as a tool to determine the specific representations of generated images and videos. While these are early results (see Figure 12 and 13), our findings indicate the methods' ability to generate realistic faces from attribute labels.