A neural network for creating images from short text descriptions with the aid of common-sense knowledge. (IMAGE)
Caption
A plan of the authors’ common-sense generative adversarial network or “CD-GAN” for text-to-image synthesis showing three modules, description enhancement, multistage generation and multistage discrimination.
Credit
Guokai Zhang et al.
Usage Restrictions
Credit must be given to the creator.
License
CC BY