Fig. 2From: Unsupervised domain adaptation for lip reading based on cross-modal knowledge distillationOverview of our proposed UDABack to article page