Self-Distillation for Unsupervised 3D Domain Adaptation

TL;DR

This paper introduces a novel self-distillation method using Graph Neural Networks for unsupervised 3D point cloud domain adaptation, achieving state-of-the-art results by enhancing discriminative features and refining pseudo-labels.

Contribution

It proposes a new iterative self-training approach with GNNs for UDA in 3D point cloud classification, improving discriminative feature learning and pseudo-label refinement.

Findings

Achieved new state-of-the-art performance on standard UDA benchmarks.

Demonstrated effectiveness of self-distillation with GNNs in unsupervised domain adaptation.

Extended the approach to complex tasks like part segmentation.

Abstract

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.