Text to Realistic Image Generation with Attentional Concatenation Generative Adversarial Networks

Figure 3

Images generated from descriptions using three GAN models trained on CUB test set.