
GAN论文学习
肉鸡小龙
这个作者很懒,什么都没留下…
展开
-
SaGAN:Generative Adversarial Network with Spatial Attention for Face Attribute Editing
#ProblemTraditional GAN methods directly operate on the whole image, and inevitably change the attribute-irrelevant regionsThe performance of traditional regression methods heavily depends on the p...原创 2019-04-20 15:26:59 · 865 阅读 · 0 评论 -
StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation
#Problemexisting models are both inefficient and ineffective in such multi-domain image translation tasksincapable of jointly training domains from different datasetsNew methodStargan, a novel ...原创 2019-04-20 15:27:40 · 590 阅读 · 0 评论