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

By Charles R. Qi et al

Published on April 10, 2017

Read the original document by opening this link in a new tab.

Table of Contents

1. Introduction
2. Related Work
3. Problem Statement
4. Deep Learning on Point Sets

Summary

PointNet is a neural network designed for processing point clouds directly. It provides a unified architecture for various 3D recognition tasks such as object classification and scene semantic parsing. The network is efficient, effective, and robust. The design choices include a symmetric function for unordered input, aggregation of local and global information, and joint alignment networks. The network architecture allows for semantic labeling of point clouds invariant to geometric transformations.