SO-Net: Self-Organizing Network for Point Cloud Analysis (2018-03-12T00:00:00.000000Z)

TL;DR

The proposed SO-Net, a permutation invariant architecture for deep learning with orderless point clouds, demonstrates performance that is similar with or better than state-of-the-art approaches in recognition tasks such as point cloud reconstruction, classification, object part segmentation and shape retrieval.

Abstract

This paper presents SO-Net, a permutation invariant architecture for deep learning with orderless point clouds. The SO-Net models the spatial distribution of point cloud by building a Self-Organizing Map (SOM). Based on the SOM, SO-Net performs hierarchical feature extraction on individual points and SOM nodes, and ultimately represents the input point cloud by a single feature vector. The receptive field of the network can be systematically adjusted by conducting point-to-node k nearest neighbor search. In recognition tasks such as point cloud reconstruction, classification, object part segmentation and shape retrieval, our proposed network demonstrates performance that is similar with or better than state-of-the-art approaches. In addition, the training speed is significantly faster than existing point cloud recognition networks because of the parallelizability and simplicity of the proposed architecture. Our code is available at the project website.1

Authors

Gim Hee Lee

9 papers

Jiaxin Li

2 papers

Ben M. Chen

1 papers

TL;DR

Abstract

Authors

References40 items

Dominant Set Clustering and Pooling for Multi-View 3D Object Recognition

Learning Representations and Generative Models for 3D Point Clouds

Locality-Sensitive Deconvolution Networks with Gated Fusion for RGB-D Indoor Semantic Segmentation

PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space

Dynamic Edge-Conditioned Filters in Convolutional Neural Networks on Graphs

Escape from Cells: Deep Kd-Networks for the Recognition of 3D Point Cloud Models

Deep Sets

Joint 2D-3D-Semantic Data for Indoor Scene Understanding

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

A Point Set Generation Network for 3D Object Reconstruction from a Single Image

OctNet: Learning Deep 3D Representations at High Resolutions

A scalable active framework for region annotation in 3D shape collections

Vote3Deep: Fast object detection in 3D point clouds using efficient convolutional neural networks

Unsupervised Monocular Depth Estimation with Left-Right Consistency

Generative and Discriminative Voxel Modeling with Convolutional Neural Networks

FPNN: Field Probing Neural Networks for 3D Data

Volumetric and Multi-view CNNs for Object Classification on 3D Data

GIFT: A Real-Time and Scalable 3D Shape Search Engine

Unsupervised CNN for Single View Depth Estimation: Geometry to the Rescue

ShapeNet: An Information-Rich 3D Model Repository

VoxNet: A 3D Convolutional Neural Network for real-time object recognition

Voting for Voting in Online Point Cloud Object Detection

Learning class‐specific descriptors for deformable shapes using localized spectral convolutional networks

Multi-view Convolutional Neural Networks for 3D Shape Recognition

Geodesic Convolutional Neural Networks on Riemannian Manifolds

Adam: A Method for Stochastic Optimization

3D ShapeNets: A deep representation for volumetric shapes

Spectral Networks and Locally Connected Networks on Graphs

Network In Network

Multi-column deep neural networks for image classification

Why Does Unsupervised Pre-training Help Deep Learning?

Best practices for convolutional neural networks applied to visual document analysis

The self-organizing map

Multidimensional binary search trees used for associative searching

Deep learning with sets and point clouds

GradientBased Learning Applied to Document Recognition

Gradient-based learning applied to document recognition

Octree encoding: A new technique for the representation, manipulation and display of arbitrary 3-d objects by computer

Supplementary A. Overview

This supplementary document provides more details and experimental results to the main paper retrieval experiments are demonstrated with

Field of Study

Journal Information

Name

Page