Self-Distillation for Unsupervised 3D Domain Adaptation
It addresses the problem of point cloud classification across different sensors or procedures for 3D vision applications, representing an incremental improvement over existing methods.
The paper tackles unsupervised domain adaptation for point cloud classification by proposing a self-distillation method that enforces consistency between a point cloud and its augmented version to obtain a discriminative feature space, achieving state-of-the-art results on standard benchmarks.
Point cloud classification is a popular task in 3D vision. However, previous works, usually assume that point clouds at test time are obtained with the same procedure or sensor as those at training time. Unsupervised Domain Adaptation (UDA) instead, breaks this assumption and tries to solve the task on an unlabeled target domain, leveraging only on a supervised source domain. For point cloud classification, recent UDA methods try to align features across domains via auxiliary tasks such as point cloud reconstruction, which however do not optimize the discriminative power in the target domain in feature space. In contrast, in this work, we focus on obtaining a discriminative feature space for the target domain enforcing consistency between a point cloud and its augmented version. We then propose a novel iterative self-training methodology that exploits Graph Neural Networks in the UDA context to refine pseudo-labels. We perform extensive experiments and set the new state-of-the-art in standard UDA benchmarks for point cloud classification. Finally, we show how our approach can be extended to more complex tasks such as part segmentation.