TY - RPRT TI - LipSound2: Self-Supervised Pre-Training for Lip-to-Speech Reconstruction and Lip Reading AU - Leyuan Qu AU - Cornelius Weber AU - Stefan Wermter PY - 2022 DO - 10.1109/tnnls.2022.3191677 UR - https://arxiv.org/abs/2112.04748 ID - 2112.04748 ER -