跳至主要內容

CycleGAN

Vingkin...小于 1 分钟

附录

图像转译和生成对抗网络GAN必读论文 pix2pix

Image-to-Image Translation with Conditional Adversarial Nets

CVPR 2017

使用条件式生成对抗网络,提出图像转译的通用框架。生成器采用U-Net网络结构,融合底层细粒度特征和高层抽象语义特征。判别器采用patchGAN网络结构,在图块尺度提取纹理等高频信息。

pix2pix在语义标签图转真实照片、简笔画转真图、黑白图像上色、卫星航拍图转地图等图像转译任务上表现优秀。

主页

论文主页:https://phillipi.github.io/pix2pix/open in new window

子豪兄论文精读视频:https://www.bilibili.com/video/BV1wY4y1k7Tc/open in new window

论文:https://arxiv.org/abs/1611.07004open in new window

代码:https://github.com/junyanz/pytorch-CycleGAN-and-pix2pixopen in new window

交互式趣味Demo:https://affinelayer.com/pixsrv/open in new window

趣味案例

床单充电线作画Gommy Sunday:https://vimeo.com/260612034open in new window

调色板生成:http://colormind.io/blog/open in new window

人脸简笔画转肖像画:https://twitter.com/quasimondo/status/826065030944870400open in new window

代码

https://colab.research.google.com/github/tensorflow/docs/blob/master/site/en/tutorials/generative/pix2pix.ipynbopen in new window

https://github.com/TommyZihao/MMGeneration_Tutorials/blob/main/【E】图像转译-pix2pix.ipynbopen in new window

评论
  • 按正序
  • 按倒序
  • 按热度
Powered by Waline v2.15.8