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

Figure 4

Self-attention mechanism introduced in the proposed GAN architecture.