Deep learning method developed for 3D reconstruction of faces
The new approach to 3D face scanning uses deep learning, auto-labeled synthetic data and unlabelled real data
The domain-adaptive reconstruction method has helped researchers to overcome training limitations that existed in previous methods due to the scarcity of real depth images with accurate 3D facial labels.
Welcome to Imaging and Machine Vision Europe
Already registered? Sign in to continue reading
Not got an account? Register for free to continue reading