PointCutMix: Regularization Strategy for Point Cloud Classification
This work addresses the problem of insufficient point cloud dataset scale and weak network generalization for researchers and practitioners working on 3D point cloud analysis, offering an incremental improvement through a novel augmentation strategy.
This paper introduces PointCutMix, a data augmentation method for point cloud classification that generates new training data by optimally assigning and replacing points between two point clouds. This method consistently and significantly improves the performance of various models and enhances model robustness against point attacks, outperforming state-of-the-art defense algorithms when used as a defense.
As 3D point cloud analysis has received increasing attention, the insufficient scale of point cloud datasets and the weak generalization ability of networks become prominent. In this paper, we propose a simple and effective augmentation method for the point cloud data, named PointCutMix, to alleviate those problems. It finds the optimal assignment between two point clouds and generates new training data by replacing the points in one sample with their optimal assigned pairs. Two replacement strategies are proposed to adapt to the accuracy or robustness requirement for different tasks, one of which is to randomly select all replacing points while the other one is to select k nearest neighbors of a single random point. Both strategies consistently and significantly improve the performance of various models on point cloud classification problems. By introducing the saliency maps to guide the selection of replacing points, the performance further improves. Moreover, PointCutMix is validated to enhance the model robustness against the point attack. It is worth noting that when using as a defense method, our method outperforms the state-of-the-art defense algorithms. The code is available at:https://github.com/cuge1995/PointCutMix