We propose an automatic character face generation method that predicts both face shape and
texture for one portrait. It can be used for most existing 3D games. Research text
In order for 3D Morphing Face Models (3DMMs) to accurately reproduce the profile of a person, they must be trained on large sets of image and texture data.
Compiling these data sets may takequite a lot of time. Also, such a system can only work stably if new data is regularly loaded. To overcome this limitation, the authors of the work, Lin, Yuan and Zou, used images of real people rather than generated photographs.
They first reconstructed the face based on3D face morphing model (3DMM) and convolutional neural networks (CNNs), and then transferring the shape of the 3D face to a grid of templates. As a result, the network receives a face image and an unrolled UV texture map as input, and then it predicts the light factors.
The authors tested their deep learning technique in a series of experiments: they compared the quality of game characters with other generated models.
Read more
Because of the Sun, the Earth's atmosphere will lose all free oxygen
Physicists have created an analogue of a black hole and confirmed Hawking's theory. Where it leads?
Abortion and science: what will happen to the children who will give birth