Core Innovation
We strive for creative yet efficient technical innovation, at the edge of AI, Computer
Graphics and Vision to push the edge of how digital humans are born.
True to life
We are committed to faithfully capture the real-world.
Our custom facial scanner is a mirror of the digital world.
Ready to work
We aim for seamless integration of technical innovation. Technical excellence is
not contradictory with ease of use and efficiency in contents creation.
Our custom head rig for facial capture is designed with the newest trends in reflectance-based geometry
reconstruction and neural inverse rendering to capture the most subtle details of the face and extract highly detailed
data ready for production.
Deep Learning-Based Unsupervised Human Facial Retargeting
S.Kim1, S.Jung1, K.Seo1, R.Blanco2, J.Noh1 1 KAIST 2 Gulliver Studios
2021, Computer Graphics Forum
This research is inspired by recent developments in face swapping and
reenactment.
We reformulate the 3D retargeting problem on the image domain with a
novel unsupervised deep learning method. This method allows for fully
unsupervised retargeting of facial expressions between models of different
configurations, and once trained, is suitable for automatic real-time
applications.
Project page (link:https://vml.kaist.ac.kr/main/international/individual/184 )