Attention-Based Image-to-Video Translation for Synthesizing Facial Expression Using GAN

Figure 9

The generated frames with MAEGAN + SA for the four categories of expression on the MUG dataset.