End-to-End Video-to-Speech Synthesis Using Generative Adversarial Networks | IEEE Journals & Magazine | IEEE Xplore