ICCV2019 ShellNet

Point Cloud

Posted by Renwu Li on 27 October 2019 · 1 min read

  • Title: 《ShellNet: Efficient Point Cloud Convolutional Neural Networks using Concentric Shells Statistics》
  • Source: ICCV2019
  • Subject: Point Cloud Convolution
  • [Paper Link]

Introduction

本文定义了一个高效的点云卷积算子,叫ShellConv,这个卷积算子和其他点云比较不一样的是使用了concentric spherical shells为邻域。

How it works

定义卷积的邻域:从一个局部点p出发,向外扩张,定义若干个环(壳shell),每个环内有n个点,环内每个点先用MLP提取高维特征,再卷积(加权求和)最大池化作为每个环的represent feature,将这些环的represent feature 做一维卷积作为当前卷积层点p的特征然后作为下一个卷积层的输入。

ShellConv operator. (a) For an input point cloud with/without associated features, representative points (red dots) are randomly sampled. The nearest neighbors are then chosen to form a point set centered at the representative points. The point sets are distributed across a series of concentric spherical shells (b) and the statistics of each shell is summarized by a maxpooling over all points in the shell, the features of which are lifted by an mlp to a higher dimension. The maxpooled features are indicated as squares with different colors (c). Following the inner to the outer order, a standard 1D convolution can be performed to yield the output features (d). Thicker dot means less points but each has higher dimensional features.

然后该文将上述所定义的卷积算子用到了点云的分类和分割上,对分割任务而言采用了U-Net的结构,将低层语义和高层语义进行连接。

Summary

本文所提出的卷积算子ShellConv相比于其他点云卷积方式,主要区别体现在了Neighborhood定义的不同,使用Shell的方式定义,从而有更大的感受野,能捕捉到local和相对不那么local的语义特征。