Adaptive Hierarchical Down-Sampling for Point Cloud Classification

04/11/2019
by   Ehsan Nezhadarya, et al.
0

While several convolution-like operators have recently been proposed for extracting features out of point clouds, down-sampling an unordered point cloud in a deep neural network has not been rigorously studied. Existing methods down-sample the points regardless of their importance for the output. As a result, some important points in the point cloud may be removed, while less valuable points may be passed to the next layers. In contrast, adaptive down-sampling methods sample the points by taking into account the importance of each point, which varies based on the application, task and training data. In this paper, we propose a permutation-invariant learning-based adaptive down-sampling layer, called Critical Points Layer (CPL), which reduces the number of points in an unordered point cloud while retaining the important points. Unlike most graph-based point cloud down-sampling methods that use k-NN search algorithm to find the neighbouring points, CPL is a global down-sampling method, rendering it computationally very efficient. The proposed layer can be used along with any graph-based point cloud convolution layer to form a convolutional neural network, dubbed CP-Net in this paper. We introduce a CP-Net for 3D object classification that achieves the best accuracy for the ModelNet40 dataset among point cloud-based methods, which validates the effectiveness of the CPL.

READ FULL TEXT
research
08/14/2018

URSA: A Neural Network for Unordered Point Clouds Using Constellations

This paper describes a neural network layer, named URSA, that uses a con...
research
05/16/2018

Critical Points to Determine Persistence Homology

Computation of the simplicial complexes of a large point cloud often rel...
research
03/23/2023

CP^3: Channel Pruning Plug-in for Point-based Networks

Channel pruning can effectively reduce both computational cost and memor...
research
12/04/2018

Learning to Sample

Processing large point clouds is a challenging task. Therefore, the data...
research
05/01/2020

MOPS-Net: A Matrix Optimization-driven Network forTask-Oriented 3D Point Cloud Downsampling

Downsampling is a commonly-used technique in 3D point cloudprocessing fo...
research
11/25/2019

Point Cloud Processing via Recurrent Set Encoding

We present a new permutation-invariant network for 3D point cloud proces...
research
06/12/2020

Local-Area-Learning Network: Meaningful Local Areas for Efficient Point Cloud Analysis

Research in point cloud analysis with deep neural networks has made rapi...

Please sign up or login with your details

Forgot password? Click here to reset