How it works
Create canonical avatar based on a three-dimension grid by Convolutional Neural Network (CNN).
Capture dense point network in the user's face by combining interpolation of logic (cubic splines) and a Convolutional Neural Network (CNN).
Mapping viewpoint mesh on avatar grid and calibration
(deep learning) for emotion depiction.
Our fully automatic and amazingly fast engine produces a lifelike 3D avatar that is ready to be animated within 5 seconds.
Start with our lightweight cloud solution scalable to unlimited number of users, or go native and integrate our engine directly into your application.