多域test时unseentarget转换-input pose 生成多条件y的外貌 few-shot unsupervised image to image translation

few-shot unsupervised image to image translation

content pose、class 外貌

target 外貌
source pose

Few-shot Unsupervised Image-to-Image Translation
• Generating images of unseen domain in the few shot setting
• Extract target domain appearance information via a class encoder with AdaIN

Paper link:
githttps://github.com/nvlabs/funit/