Deep generative image models using a laplacian pyramid of adversarial networks

Emily Denton, Soumith Chintala, Arthur Szlam, Robert Fergus

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this paper we introduce a generative parametric model capable of producing high quality samples of natural images. Our approach uses a cascade of convolutional networks within a Laplacian pyramid framework to generate images in a coarse-to-fine fashion. At each level of the pyramid, a separate generative convnet model is trained using the Generative Adversarial Nets (GAN) approach [11]. Samples drawn from our model are of significantly higher quality than alternate approaches. In a quantitative assessment by human evaluators, our CIFAR10 samples were mistaken for real images around 40% of the time, compared to 10% for samples drawn from a GAN baseline model. We also show samples from models trained on the higher resolution images of the LSUN scene dataset.

Original languageEnglish (US)
Title of host publicationAdvances in Neural Information Processing Systems
PublisherNeural information processing systems foundation
Pages1486-1494
Number of pages9
Volume2015-January
StatePublished - 2015
Event29th Annual Conference on Neural Information Processing Systems, NIPS 2015 - Montreal, Canada
Duration: Dec 7 2015Dec 12 2015

Other

Other29th Annual Conference on Neural Information Processing Systems, NIPS 2015
CountryCanada
CityMontreal
Period12/7/1512/12/15

Fingerprint

Image resolution

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Information Systems
  • Signal Processing

Cite this

Denton, E., Chintala, S., Szlam, A., & Fergus, R. (2015). Deep generative image models using a laplacian pyramid of adversarial networks. In Advances in Neural Information Processing Systems (Vol. 2015-January, pp. 1486-1494). Neural information processing systems foundation.

Deep generative image models using a laplacian pyramid of adversarial networks. / Denton, Emily; Chintala, Soumith; Szlam, Arthur; Fergus, Robert.

Advances in Neural Information Processing Systems. Vol. 2015-January Neural information processing systems foundation, 2015. p. 1486-1494.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Denton, E, Chintala, S, Szlam, A & Fergus, R 2015, Deep generative image models using a laplacian pyramid of adversarial networks. in Advances in Neural Information Processing Systems. vol. 2015-January, Neural information processing systems foundation, pp. 1486-1494, 29th Annual Conference on Neural Information Processing Systems, NIPS 2015, Montreal, Canada, 12/7/15.
Denton E, Chintala S, Szlam A, Fergus R. Deep generative image models using a laplacian pyramid of adversarial networks. In Advances in Neural Information Processing Systems. Vol. 2015-January. Neural information processing systems foundation. 2015. p. 1486-1494
Denton, Emily ; Chintala, Soumith ; Szlam, Arthur ; Fergus, Robert. / Deep generative image models using a laplacian pyramid of adversarial networks. Advances in Neural Information Processing Systems. Vol. 2015-January Neural information processing systems foundation, 2015. pp. 1486-1494
@inproceedings{1fcb4b35451a4c5c9f42725b0ba71437,
title = "Deep generative image models using a laplacian pyramid of adversarial networks",
abstract = "In this paper we introduce a generative parametric model capable of producing high quality samples of natural images. Our approach uses a cascade of convolutional networks within a Laplacian pyramid framework to generate images in a coarse-to-fine fashion. At each level of the pyramid, a separate generative convnet model is trained using the Generative Adversarial Nets (GAN) approach [11]. Samples drawn from our model are of significantly higher quality than alternate approaches. In a quantitative assessment by human evaluators, our CIFAR10 samples were mistaken for real images around 40{\%} of the time, compared to 10{\%} for samples drawn from a GAN baseline model. We also show samples from models trained on the higher resolution images of the LSUN scene dataset.",
author = "Emily Denton and Soumith Chintala and Arthur Szlam and Robert Fergus",
year = "2015",
language = "English (US)",
volume = "2015-January",
pages = "1486--1494",
booktitle = "Advances in Neural Information Processing Systems",
publisher = "Neural information processing systems foundation",

}

TY - GEN

T1 - Deep generative image models using a laplacian pyramid of adversarial networks

AU - Denton, Emily

AU - Chintala, Soumith

AU - Szlam, Arthur

AU - Fergus, Robert

PY - 2015

Y1 - 2015

N2 - In this paper we introduce a generative parametric model capable of producing high quality samples of natural images. Our approach uses a cascade of convolutional networks within a Laplacian pyramid framework to generate images in a coarse-to-fine fashion. At each level of the pyramid, a separate generative convnet model is trained using the Generative Adversarial Nets (GAN) approach [11]. Samples drawn from our model are of significantly higher quality than alternate approaches. In a quantitative assessment by human evaluators, our CIFAR10 samples were mistaken for real images around 40% of the time, compared to 10% for samples drawn from a GAN baseline model. We also show samples from models trained on the higher resolution images of the LSUN scene dataset.

AB - In this paper we introduce a generative parametric model capable of producing high quality samples of natural images. Our approach uses a cascade of convolutional networks within a Laplacian pyramid framework to generate images in a coarse-to-fine fashion. At each level of the pyramid, a separate generative convnet model is trained using the Generative Adversarial Nets (GAN) approach [11]. Samples drawn from our model are of significantly higher quality than alternate approaches. In a quantitative assessment by human evaluators, our CIFAR10 samples were mistaken for real images around 40% of the time, compared to 10% for samples drawn from a GAN baseline model. We also show samples from models trained on the higher resolution images of the LSUN scene dataset.

UR - http://www.scopus.com/inward/record.url?scp=84965143571&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84965143571&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84965143571

VL - 2015-January

SP - 1486

EP - 1494

BT - Advances in Neural Information Processing Systems

PB - Neural information processing systems foundation

ER -