Point Cloud Analysis Under Slight Perturbations: A Manifold Distillation Approach Using Raw Coordinates

Point cloud is often regarded as a discrete sampling of Riemannian manifold and plays a pivotal role in the 3D image interpretation. Particularly, rotation perturbation, an unexpected small change in rotation caused by various factors (like equipment offset, system instability, measurement errors and so on), can easily lead to the inferior results in point cloud learning tasks. However, classical point cloud learning methods are sensitive to rotation perturbation, and the existing networks with rotation robustness also have much room for improvements in terms of performance and noise tolerance. Given these, this paper remodels the point cloud from the perspective of manifold as well as designs a manifold distillation method to achieve the robustness of rotation perturbation without any coordinate transformation. In brief, during the training phase, we introduce a teacher network to learn the rotation robustness information and transfer this information to the student network through online distillation. In the inference phase, the student network directly utilizes the raw 3D coordinate information to achieve the robustness of rotation perturbation. Experiments carried out on four different datasets verify the effectiveness of our method. On average, on the ModelNet40 and ScanObjectNN classification datasets with random rotation perturbations, our method improves classification accuracy by 4.41% and 3.65%, respectively, compared to popular rotation-robust networks. Similarly, on the ShapeNet and S3DIS segmentation datasets, our method achieves improvements in mIoU of 6.96% and 5.12%, respectively. Furthermore, the experimental results also demonstrate that our algorithm exhibits higher computational efficiency and stronger resistance to noise and outliers.