Abstract. In this paper, we introduce a novel unsupervised domain adaptation technique for the task of 3D keypoint prediction from a single depth scan or image. Our key idea is to utilize the fact that predictions from difffferent views of the same or similar objects should be consistent with each other. Such view consistency can provide effffective regularization for keypoint prediction on unlabeled instances. In addition, we introduce a geometric alignment term to regularize predictions in the target domain. The resulting loss function can be effffectively optimized via alternating minimization. We demonstrate the effffectiveness of our approach on real datasets and present experimental results showing that our approach is superior to state-of-the-art general-purpose domain adaptation techniques