Check out a new advanced framework proposed by a team from Shanghai Jiao Tong University.
Have a look at a new advanced face-swapping framework called Simple Swap (SimSwap). Unlike previous approaches, the new model allows users to keep original attributes like makeup, face painting, and more.
The team behind the research states that they achieved the result with two main steps. First, they presented the ID Injection Module (IIM) to transfer the identity information of the source face at the feature level. They have also introduced the Weak Feature Matching Loss to preserve the facial attributes in an implicit way.
You can learn more and find the code on GitHub. Don't forget to join our new Telegram channel, our Discord, follow us on Instagram and Twitter, where we are sharing breakdowns, the latest news, awesome artworks, and more.