Pioneer Networks: Progressively Growing Generative Autoencoder

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussavertaisarvioitu

Standard

Pioneer Networks : Progressively Growing Generative Autoencoder. / Heljakka, Ari; Solin, Arno; Kannala, Juho.

Computer Vision – ACCV 2018 - 14th Asian Conference on Computer Vision, Revised Selected Papers. toim. / Greg Mori; C.V. Jawahar; Konrad Schindler; Hongdong Li. Springer, Cham, 2019. s. 22-38 (Lecture notes in computer science; Vuosikerta 11361).

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussavertaisarvioitu

Harvard

Heljakka, A, Solin, A & Kannala, J 2019, Pioneer Networks: Progressively Growing Generative Autoencoder. julkaisussa G Mori, CV Jawahar, K Schindler & H Li (toim), Computer Vision – ACCV 2018 - 14th Asian Conference on Computer Vision, Revised Selected Papers. Lecture notes in computer science, Vuosikerta. 11361, Springer, Cham, Sivut 22-38, Asian Conference on Computer Vision , Perth, Austraalia, 02/12/2018. https://doi.org/10.1007/978-3-030-20887-5_2

APA

Heljakka, A., Solin, A., & Kannala, J. (2019). Pioneer Networks: Progressively Growing Generative Autoencoder. teoksessa G. Mori, C. V. Jawahar, K. Schindler, & H. Li (Toimittajat), Computer Vision – ACCV 2018 - 14th Asian Conference on Computer Vision, Revised Selected Papers (Sivut 22-38). (Lecture notes in computer science; Vuosikerta 11361). Springer, Cham. https://doi.org/10.1007/978-3-030-20887-5_2

Vancouver

Heljakka A, Solin A, Kannala J. Pioneer Networks: Progressively Growing Generative Autoencoder. julkaisussa Mori G, Jawahar CV, Schindler K, Li H, toimittajat, Computer Vision – ACCV 2018 - 14th Asian Conference on Computer Vision, Revised Selected Papers. Springer, Cham. 2019. s. 22-38. (Lecture notes in computer science). https://doi.org/10.1007/978-3-030-20887-5_2

Author

Heljakka, Ari ; Solin, Arno ; Kannala, Juho. / Pioneer Networks : Progressively Growing Generative Autoencoder. Computer Vision – ACCV 2018 - 14th Asian Conference on Computer Vision, Revised Selected Papers. Toimittaja / Greg Mori ; C.V. Jawahar ; Konrad Schindler ; Hongdong Li. Springer, Cham, 2019. Sivut 22-38 (Lecture notes in computer science).

Bibtex - Lataa

@inproceedings{68b6dc32099f49d4996ff47af6001c0e,
title = "Pioneer Networks: Progressively Growing Generative Autoencoder",
abstract = "We introduce a novel generative autoencoder network model that learns to encode and reconstruct images with high quality and resolution, and supports smooth random sampling from the latent space of the encoder. Generative adversarial networks (GANs) are known for their ability to simulate random high-quality images, but they cannot reconstruct existing images. Previous works have attempted to extend GANs to support such inference but, so far, have not delivered satisfactory high-quality results. Instead, we propose the Progressively Growing Generative Autoencoder (Pioneer) network which achieves high-quality reconstruction with images without requiring a GAN discriminator. We merge recent techniques for progressively building up the parts of the network with the recently introduced adversarial encoder–generator network. The ability to reconstruct input images is crucial in many real-world applications, and allows for precise intelligent manipulation of existing images. We show promising results in image synthesis and inference, with state-of-the-art results in CelebA inference tasks.",
keywords = "Autoencoder, Computer vision, Generative models",
author = "Ari Heljakka and Arno Solin and Juho Kannala",
year = "2019",
doi = "10.1007/978-3-030-20887-5_2",
language = "English",
isbn = "978-3-030-20886-8",
series = "Lecture notes in computer science",
publisher = "Springer Nature",
pages = "22--38",
editor = "Greg Mori and C.V. Jawahar and Konrad Schindler and Hongdong Li",
booktitle = "Computer Vision – ACCV 2018 - 14th Asian Conference on Computer Vision, Revised Selected Papers",

}

RIS - Lataa

TY - GEN

T1 - Pioneer Networks

T2 - Progressively Growing Generative Autoencoder

AU - Heljakka, Ari

AU - Solin, Arno

AU - Kannala, Juho

PY - 2019

Y1 - 2019

N2 - We introduce a novel generative autoencoder network model that learns to encode and reconstruct images with high quality and resolution, and supports smooth random sampling from the latent space of the encoder. Generative adversarial networks (GANs) are known for their ability to simulate random high-quality images, but they cannot reconstruct existing images. Previous works have attempted to extend GANs to support such inference but, so far, have not delivered satisfactory high-quality results. Instead, we propose the Progressively Growing Generative Autoencoder (Pioneer) network which achieves high-quality reconstruction with images without requiring a GAN discriminator. We merge recent techniques for progressively building up the parts of the network with the recently introduced adversarial encoder–generator network. The ability to reconstruct input images is crucial in many real-world applications, and allows for precise intelligent manipulation of existing images. We show promising results in image synthesis and inference, with state-of-the-art results in CelebA inference tasks.

AB - We introduce a novel generative autoencoder network model that learns to encode and reconstruct images with high quality and resolution, and supports smooth random sampling from the latent space of the encoder. Generative adversarial networks (GANs) are known for their ability to simulate random high-quality images, but they cannot reconstruct existing images. Previous works have attempted to extend GANs to support such inference but, so far, have not delivered satisfactory high-quality results. Instead, we propose the Progressively Growing Generative Autoencoder (Pioneer) network which achieves high-quality reconstruction with images without requiring a GAN discriminator. We merge recent techniques for progressively building up the parts of the network with the recently introduced adversarial encoder–generator network. The ability to reconstruct input images is crucial in many real-world applications, and allows for precise intelligent manipulation of existing images. We show promising results in image synthesis and inference, with state-of-the-art results in CelebA inference tasks.

KW - Autoencoder

KW - Computer vision

KW - Generative models

UR - http://www.scopus.com/inward/record.url?scp=85066805190&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-20887-5_2

DO - 10.1007/978-3-030-20887-5_2

M3 - Conference contribution

SN - 978-3-030-20886-8

T3 - Lecture notes in computer science

SP - 22

EP - 38

BT - Computer Vision – ACCV 2018 - 14th Asian Conference on Computer Vision, Revised Selected Papers

A2 - Mori, Greg

A2 - Jawahar, C.V.

A2 - Schindler, Konrad

A2 - Li, Hongdong

CY - Springer, Cham

ER -

ID: 33778897