Publications

Exchangeable Feature Alignment for Arbitrary Style Transfer

ETNet: Error Transition Network for Arbitrary Style Transfer

Projects

Global-to-Local Generative Model for 3D Shapes

We introduce a novel global-to-local(G2L) approach for 3D model generation through building an adversarial network(GAN) assiciated with additional local discriminators to construct the overall structure of the shape while yielding more plausible part geometries.

Contact

  • songchunjin1990[at]gmail[dot]com
  • Shenzhen University South Campus L6 - 812, Guangdong Province, China