Self-Supervised Learning for Domain Adaptation on Point-Clouds
This addresses domain adaptation challenges in 3D perception for applications like robotics and autonomous driving, but it is incremental as it extends known techniques to a new modality.
The paper tackled the problem of applying self-supervised learning to domain adaptation for 3D point clouds, introducing a new pretext task and training procedure that demonstrated large improvements over existing methods on classification and segmentation datasets.
Self-supervised learning (SSL) is a technique for learning useful representations from unlabeled data. It has been applied effectively to domain adaptation (DA) on images and videos. It is still unknown if and how it can be leveraged for domain adaptation in 3D perception problems. Here we describe the first study of SSL for DA on point clouds. We introduce a new family of pretext tasks, Deformation Reconstruction, inspired by the deformations encountered in sim-to-real transformations. In addition, we propose a novel training procedure for labeled point cloud data motivated by the MixUp method called Point cloud Mixup (PCM). Evaluations on domain adaptations datasets for classification and segmentation, demonstrate a large improvement over existing and baseline methods.