TUTCRIS - Tampereen teknillinen yliopisto

TUTCRIS

Predicting Novel Views Using Generative Adversarial Query Network

Tutkimustuotosvertaisarvioitu

Standard

Predicting Novel Views Using Generative Adversarial Query Network. / Nguyen-Ha, Phong; Huynh, Lam; Rahtu, Esa; Heikkilä, Janne.

Image Analysis - 21st Scandinavian Conference, SCIA 2019, Proceedings. toim. / Michael Felsberg; Per-Erik Forssén; Jonas Unger; Ida-Maria Sintorn. Springer Verlag, 2019. s. 16-27 (Lecture Notes in Computer Science; Vuosikerta 11482).

Tutkimustuotosvertaisarvioitu

Harvard

Nguyen-Ha, P, Huynh, L, Rahtu, E & Heikkilä, J 2019, Predicting Novel Views Using Generative Adversarial Query Network. julkaisussa M Felsberg, P-E Forssén, J Unger & I-M Sintorn (toim), Image Analysis - 21st Scandinavian Conference, SCIA 2019, Proceedings. Lecture Notes in Computer Science, Vuosikerta. 11482, Springer Verlag, Sivut 16-27, Norrköping, Ruotsi, 11/06/19. https://doi.org/10.1007/978-3-030-20205-7_2

APA

Nguyen-Ha, P., Huynh, L., Rahtu, E., & Heikkilä, J. (2019). Predicting Novel Views Using Generative Adversarial Query Network. teoksessa M. Felsberg, P-E. Forssén, J. Unger, & I-M. Sintorn (Toimittajat), Image Analysis - 21st Scandinavian Conference, SCIA 2019, Proceedings (Sivut 16-27). (Lecture Notes in Computer Science; Vuosikerta 11482). Springer Verlag. https://doi.org/10.1007/978-3-030-20205-7_2

Vancouver

Nguyen-Ha P, Huynh L, Rahtu E, Heikkilä J. Predicting Novel Views Using Generative Adversarial Query Network. julkaisussa Felsberg M, Forssén P-E, Unger J, Sintorn I-M, toimittajat, Image Analysis - 21st Scandinavian Conference, SCIA 2019, Proceedings. Springer Verlag. 2019. s. 16-27. (Lecture Notes in Computer Science). https://doi.org/10.1007/978-3-030-20205-7_2

Author

Nguyen-Ha, Phong ; Huynh, Lam ; Rahtu, Esa ; Heikkilä, Janne. / Predicting Novel Views Using Generative Adversarial Query Network. Image Analysis - 21st Scandinavian Conference, SCIA 2019, Proceedings. Toimittaja / Michael Felsberg ; Per-Erik Forssén ; Jonas Unger ; Ida-Maria Sintorn. Springer Verlag, 2019. Sivut 16-27 (Lecture Notes in Computer Science).

Bibtex - Lataa

@inproceedings{3cf183b9ae124544890d0e2a9047dadd,
title = "Predicting Novel Views Using Generative Adversarial Query Network",
abstract = "The problem of predicting a novel view of the scene using an arbitrary number of observations is a challenging problem for computers as well as for humans. This paper introduces the Generative Adversarial Query Network (GAQN), a general learning framework for novel view synthesis that combines Generative Query Network (GQN) and Generative Adversarial Networks (GANs). The conventional GQN encodes input views into a latent representation that is used to generate a new view through a recurrent variational decoder. The proposed GAQN builds on this work by adding two novel aspects: First, we extend the current GQN architecture with an adversarial loss function for improving the visual quality and convergence speed. Second, we introduce a feature-matching loss function for stabilizing the training procedure. The experiments demonstrate that GAQN is able to produce high-quality results and faster convergence compared to the conventional approach.",
keywords = "Generative Adversarial Query Network, Mean feature matching loss, Novel view synthesis",
author = "Phong Nguyen-Ha and Lam Huynh and Esa Rahtu and Janne Heikkil{\"a}",
note = "jufoid=62555",
year = "2019",
doi = "10.1007/978-3-030-20205-7_2",
language = "English",
isbn = "9783030202040",
series = "Lecture Notes in Computer Science",
publisher = "Springer Verlag",
pages = "16--27",
editor = "Michael Felsberg and Per-Erik Forss{\'e}n and Jonas Unger and Ida-Maria Sintorn",
booktitle = "Image Analysis - 21st Scandinavian Conference, SCIA 2019, Proceedings",
address = "Germany",

}

RIS (suitable for import to EndNote) - Lataa

TY - GEN

T1 - Predicting Novel Views Using Generative Adversarial Query Network

AU - Nguyen-Ha, Phong

AU - Huynh, Lam

AU - Rahtu, Esa

AU - Heikkilä, Janne

N1 - jufoid=62555

PY - 2019

Y1 - 2019

N2 - The problem of predicting a novel view of the scene using an arbitrary number of observations is a challenging problem for computers as well as for humans. This paper introduces the Generative Adversarial Query Network (GAQN), a general learning framework for novel view synthesis that combines Generative Query Network (GQN) and Generative Adversarial Networks (GANs). The conventional GQN encodes input views into a latent representation that is used to generate a new view through a recurrent variational decoder. The proposed GAQN builds on this work by adding two novel aspects: First, we extend the current GQN architecture with an adversarial loss function for improving the visual quality and convergence speed. Second, we introduce a feature-matching loss function for stabilizing the training procedure. The experiments demonstrate that GAQN is able to produce high-quality results and faster convergence compared to the conventional approach.

AB - The problem of predicting a novel view of the scene using an arbitrary number of observations is a challenging problem for computers as well as for humans. This paper introduces the Generative Adversarial Query Network (GAQN), a general learning framework for novel view synthesis that combines Generative Query Network (GQN) and Generative Adversarial Networks (GANs). The conventional GQN encodes input views into a latent representation that is used to generate a new view through a recurrent variational decoder. The proposed GAQN builds on this work by adding two novel aspects: First, we extend the current GQN architecture with an adversarial loss function for improving the visual quality and convergence speed. Second, we introduce a feature-matching loss function for stabilizing the training procedure. The experiments demonstrate that GAQN is able to produce high-quality results and faster convergence compared to the conventional approach.

KW - Generative Adversarial Query Network

KW - Mean feature matching loss

KW - Novel view synthesis

U2 - 10.1007/978-3-030-20205-7_2

DO - 10.1007/978-3-030-20205-7_2

M3 - Conference contribution

SN - 9783030202040

T3 - Lecture Notes in Computer Science

SP - 16

EP - 27

BT - Image Analysis - 21st Scandinavian Conference, SCIA 2019, Proceedings

A2 - Felsberg, Michael

A2 - Forssén, Per-Erik

A2 - Unger, Jonas

A2 - Sintorn, Ida-Maria

PB - Springer Verlag

ER -