基于条件的边界平衡生成对抗网络

Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2019-04-01 Cooperative journals: 《计算机应用研究》

王硕诚 苟刚 葛梦园

Abstract： Generative adversarial networks (GAN) is one of the most popular generation models of the year. Using the generation antagonistic network and some of its improved models, the model can generate random images, or specific images of low quality. At present, there is no generation model that can use simple network structure to generate high-quality specific images. For this task, the method combines the advantages of boundary equilibrium generative adversarial network(BEGAN) , adds additional condition features and the MSE loss and establishes the conditional boundary equilibrium generative adversarial network(C-BEGAN) . This method used to extract the generation model for specific image generation. Experimental results show that compared with other supervised class generation models, this method can use simpler networks to achieve faster convergence speed and generate images with better quality and diversity.

 Hits 10150  Downloads 1032  Comment 0