New GAN Can Lipread and Synthesize Speech

Researchers from Samsung and Imperial College in London developed a deep learning solution that uses computer vision for visual speech recognition.