Text-to-image Generation Based on Attention Mechanisms
With the prevalence of artificial intelligence,text-generated images are gaining more and more attention,and it is even more important to generate higher quality images.In this paper we study the generative adversarial network model based on the attention mechanism,introduce the SimAM attention mechanism on the basis of the AttnGAN model,and improve the model architecture,so that the model can reduce the number of parameters under the premise of improving the quality of generated images.Experimental results show that the image quality generated by the improved model is indeed improved.
万方数据
维普
