microsoft/CoCosNet
A CVPR 2020 paper implementing cross-domain correspondence learning for exemplar-based image translation using PyTorch and GANs.
Not currently ranked — collecting fresh signals.
star history
CoCosNet is a deep learning framework that synthesizes photo-realistic images from input images in one domain (e.g., segmentation masks, edge maps, pose keypoints) given an exemplar image whose style it should match. It jointly learns cross-domain correspondence and image translation by first aligning images from distinct domains to an intermediate domain where dense correspondence is established, then synthesizing images based on semantically corresponding patches in the exemplar.
Frequently asked
- What is microsoft/CoCosNet?
- A CVPR 2020 paper implementing cross-domain correspondence learning for exemplar-based image translation using PyTorch and GANs.
- Is CoCosNet open source?
- Yes — microsoft/CoCosNet is open source, released under the MIT license.
- What language is CoCosNet written in?
- microsoft/CoCosNet is primarily written in Python.
- How popular is CoCosNet?
- microsoft/CoCosNet has 399 stars on GitHub.
- Where can I find CoCosNet?
- microsoft/CoCosNet is on GitHub at https://github.com/microsoft/CoCosNet.