Speaker：Professor Shang-Hong Lai（Department of Computer Science, National Tsing Hua University）
Theme：GAN models for visual learning
Time：12/19(Tuesday)PM 13：30
Venue：GEN III Room 837
Abstract：Generative Adversarial Network (GAN) has attracted a considerable amount of attention in deep learning and computer vision communities. In this talk, I will present three different GAN models for solving the image completion, face frontalization, and day-to-night image conversion problems, respectively. Each GAN model is specifically designed for the corresponding problem and we propose an associated loss function for the training. Some experimental results on real data will be shown to demonstrate superior performance by using the proposed models.
＜＜Welcome teachers, students enthusiastically involved！！＞＞