PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation

TLDR

Point clouds are a key geometric data structure, yet their irregular format forces most researchers to convert them to regular voxel grids or image collections, increasing volume and causing problems. The paper introduces a novel neural network that directly consumes point clouds while respecting permutation invariance. The network processes point clouds through a permutation‑invariant architecture and is theoretically analyzed for robustness to input perturbation and corruption. PointNet unifies object classification, part segmentation, and scene parsing, achieving high efficiency and performance that matches or surpasses state‑of‑the‑art methods.

Abstract

Point cloud is an important type of geometric data structure. Due to its irregular format, most researchers transform such data to regular 3D voxel grids or collections of images. This, however, renders data unnecessarily voluminous and causes issues. In this paper, we design a novel type of neural network that directly consumes point clouds, which well respects the permutation invariance of points in the input. Our network, named PointNet, provides a unified architecture for applications ranging from object classification, part segmentation, to scene semantic parsing. Though simple, PointNet is highly efficient and effective. Empirically, it shows strong performance on par or even better than state of the art. Theoretically, we provide analysis towards understanding of what the network has learnt and why the network is robust with respect to input perturbation and corruption.