SampleNet: Differentiable Point Cloud Sampling

TL;DR

SampleNet introduces a differentiable point cloud sampling method that improves downstream tasks like classification, reconstruction, and registration by enabling task-specific learning of the sampling process.

Contribution

It proposes a novel differentiable relaxation for point cloud sampling, allowing end-to-end learning and outperforming existing methods.

Findings

Consistently good results on classification tasks

Effective in geometry reconstruction applications

Outperforms existing sampling methods in various tasks

Abstract

There is a growing number of tasks that work directly on point clouds. As the size of the point cloud grows, so do the computational demands of these tasks. A possible solution is to sample the point cloud first. Classic sampling approaches, such as farthest point sampling (FPS), do not consider the downstream task. A recent work showed that learning a task-specific sampling can improve results significantly. However, the proposed technique did not deal with the non-differentiability of the sampling operation and offered a workaround instead. We introduce a novel differentiable relaxation for point cloud sampling that approximates sampled points as a mixture of points in the primary input cloud. Our approximation scheme leads to consistently good results on classification and geometry reconstruction applications. We also show that the proposed sampling method can be used as a front to a point cloud registration network. This is a challenging task since sampling must be consistent across two different point clouds for a shared downstream task. In all cases, our approach outperforms existing non-learned and learned sampling alternatives. Our code is publicly available at https://github.com/itailang/SampleNet.