Curvature Diversity-Driven Deformation and Domain Alignment for Point Cloud

Unsupervised Domain Adaptation (UDA) is crucial for reducing the need for extensive manual data annotation when training deep networks on point cloud data. A significant challenge of UDA lies in effectively bridging the domain gap. To tackle this challenge, we propose Curvature Diversity-Driven Nuclear-Norm Wasserstein Domain Alignment (CDND). Our approach first introduces a textitCurvature Diversity-driven Deformation Reconstruction (CurvRec) task, which effectively mitigates the gap between the source and target domains by enabling the model to extract salient features from semantically rich regions of a given point cloud. We then propose textitDeformation-based Nuclear-norm Wasserstein Discrepancy (D-NWD), which applies the Nuclear-norm Wasserstein Discrepancy to both textitdeformed and original data samples to align the source and target domains. Furthermore, we contribute a theoretical justification for the effectiveness of D-NWD in distribution alignment and demonstrate that it is textitgeneric enough to be applied to any deformations. To validate our method, we conduct extensive experiments on two public domain adaptation datasets for point cloud classification and segmentation tasks. Empirical experiment results show that our CDND achieves state-of-the-art performance by a noticeable margin over existing approaches.