Generative and Discriminative Voxel Modeling with Convolutional Neural Networks

Andrew Brock; Theodore Lim; James Millar Ritchie; Nicholas J. Weston

Generative and Discriminative Voxel Modeling with Convolutional Neural Networks

Andrew Brock
, Theodore Lim
, James Millar Ritchie
, Nicholas J. Weston

Research output: Contribution to conference › Other › peer-review

262 Downloads (Pure)

Abstract

When working with three-dimensional data, choice of representation is key. We
explore voxel-based models, and present evidence for the viability of voxellated
representations in applications including shape modeling and object classification. Our key contributions are methods for training voxel-based variational autoencoders, a user interface for exploring the latent space learned by the autoencoder, and a deep convolutional neural network architecture for object classification. We address challenges unique to voxel-based representations, and empirically evaluate our models on the ModelNet benchmark, where we demonstrate a 51.5% relative improvement in the state of the art for object classification.

Original language	English
Pages	1-9
Number of pages	9
Publication status	Published - 5 Dec 2016
Event	Neural Inofrmation Processing Conference: 3D Deep Learning - Barcelona, Barcelona, Spain Duration: 5 Dec 2016 → 10 Dec 2016 https://nips.cc/Conferences/2016

Conference

Conference	Neural Inofrmation Processing Conference
Abbreviated title	NIPS
Country/Territory	Spain
City	Barcelona
Period	5/12/16 → 10/12/16
Internet address	https://nips.cc/Conferences/2016

Keywords

ModelNet benchmark
3D Data
Convolutional Neural Networks
Variational Autoencoders

Access to Document

NIPS_paper_2016Final published version, 635 KBLicence: CC BY

http://3ddl.cs.princeton.edu/2016/papers/Brock_et_al.pdfLicence: CC BY

Theodore Lim
- T.Limhw.acuk
- School of Engineering & Physical Sciences - Associate Professor
- School of Engineering & Physical Sciences, Institute of Mechanical, Process & Energy Engineering - Associate Professor
Person: Academic (Research & Teaching)

Cite this

@conference{de3f5977da1342faa8d4ff92cdbf1529,

title = "Generative and Discriminative Voxel Modeling with Convolutional Neural Networks",

abstract = "When working with three-dimensional data, choice of representation is key. We explore voxel-based models, and present evidence for the viability of voxellated representations in applications including shape modeling and object classification. Our key contributions are methods for training voxel-based variational autoencoders, a user interface for exploring the latent space learned by the autoencoder, and a deep convolutional neural network architecture for object classification. We address challenges unique to voxel-based representations, and empirically evaluate our models on the ModelNet benchmark, where we demonstrate a 51.5\% relative improvement in the state of the art for object classification.",

keywords = "ModelNet benchmark, 3D Data, Convolutional Neural Networks, Variational Autoencoders",

author = "Andrew Brock and Theodore Lim and Ritchie, \{James Millar\} and Weston, \{Nicholas J.\}",

note = "Workshop contribution; Neural Inofrmation Processing Conference : 3D Deep Learning, NIPS ; Conference date: 05-12-2016 Through 10-12-2016",

year = "2016",

month = dec,

day = "5",

language = "English",

pages = "1--9",

url = "https://nips.cc/Conferences/2016",

}

TY - CONF

T1 - Generative and Discriminative Voxel Modeling with Convolutional Neural Networks

AU - Brock, Andrew

AU - Lim, Theodore

AU - Ritchie, James Millar

AU - Weston, Nicholas J.

N1 - Workshop contribution

PY - 2016/12/5

Y1 - 2016/12/5

N2 - When working with three-dimensional data, choice of representation is key. We explore voxel-based models, and present evidence for the viability of voxellated representations in applications including shape modeling and object classification. Our key contributions are methods for training voxel-based variational autoencoders, a user interface for exploring the latent space learned by the autoencoder, and a deep convolutional neural network architecture for object classification. We address challenges unique to voxel-based representations, and empirically evaluate our models on the ModelNet benchmark, where we demonstrate a 51.5% relative improvement in the state of the art for object classification.

AB - When working with three-dimensional data, choice of representation is key. We explore voxel-based models, and present evidence for the viability of voxellated representations in applications including shape modeling and object classification. Our key contributions are methods for training voxel-based variational autoencoders, a user interface for exploring the latent space learned by the autoencoder, and a deep convolutional neural network architecture for object classification. We address challenges unique to voxel-based representations, and empirically evaluate our models on the ModelNet benchmark, where we demonstrate a 51.5% relative improvement in the state of the art for object classification.

KW - ModelNet benchmark

KW - 3D Data

KW - Convolutional Neural Networks

KW - Variational Autoencoders

M3 - Other

SP - 1

EP - 9

T2 - Neural Inofrmation Processing Conference

Y2 - 5 December 2016 through 10 December 2016

ER -

Generative and Discriminative Voxel Modeling with Convolutional Neural Networks

Abstract

Conference

Keywords

Access to Document

Fingerprint

Profiles

Theodore Lim

Cite this