Danişman Aǧ Kullanarak Derin Otomatik Kodlayicilar ile İçboyama

Ugur Demir; Gozde Unal

doi:10.1109/SIU.2017.7960317

Danişman Aǧ Kullanarak Derin Otomatik Kodlayicilar ile İçboyama

Translated title of the contribution: Inpainting by deep autoencoders using an advisor network

Ugur Demir^*, Gozde Unal

^*Corresponding author for this work

Department of Artificial Intelligence and Data Engineering

Istanbul Technical University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

6 Citations (Scopus)

Abstract

In this study, we tried to find a solution for inpainting problem using deep convolutional autoencoders. A new training approach has been proposed as an alternative to the Generative Adversarial Networks. The neural network that designed for inpainting takes an image, which the certain part of its center is extracted, as an input then it attempts to fill the blank region. During the training phase, a distinct deep convolutional neural network is used and it is called Advisor Network. We show that the features extracted from intermediate layers of the Advisor Network, which is trained on a different dataset for classification, improves the performance of the autoencoder.

Translated title of the contribution	Inpainting by deep autoencoders using an advisor network
Original language	Turkish
Title of host publication	2017 25th Signal Processing and Communications Applications Conference, SIU 2017
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9781509064946
DOIs	https://doi.org/10.1109/SIU.2017.7960317
Publication status	Published - 27 Jun 2017
Event	25th Signal Processing and Communications Applications Conference, SIU 2017 - Antalya, Turkey Duration: 15 May 2017 → 18 May 2017

Publication series

Name	2017 25th Signal Processing and Communications Applications Conference, SIU 2017

Conference

Conference	25th Signal Processing and Communications Applications Conference, SIU 2017
Country/Territory	Turkey
City	Antalya
Period	15/05/17 → 18/05/17

Bibliographical note

Publisher Copyright:
© 2017 IEEE.

Access to Document

10.1109/SIU.2017.7960317

Cite this

Demir, U., & Unal, G. (2017). Danişman Aǧ Kullanarak Derin Otomatik Kodlayicilar ile İçboyama. In 2017 25th Signal Processing and Communications Applications Conference, SIU 2017 Article 7960317 (2017 25th Signal Processing and Communications Applications Conference, SIU 2017). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/SIU.2017.7960317

@inproceedings{c4802e61917547f7899d61db26e98bef,

title = "Dani{\c s}man Aǧ Kullanarak Derin Otomatik Kodlayicilar ile İ{\c c}boyama",

abstract = "In this study, we tried to find a solution for inpainting problem using deep convolutional autoencoders. A new training approach has been proposed as an alternative to the Generative Adversarial Networks. The neural network that designed for inpainting takes an image, which the certain part of its center is extracted, as an input then it attempts to fill the blank region. During the training phase, a distinct deep convolutional neural network is used and it is called Advisor Network. We show that the features extracted from intermediate layers of the Advisor Network, which is trained on a different dataset for classification, improves the performance of the autoencoder.",

keywords = "auto-encoder, convolutional neural network, deep learning, inpainting",

author = "Ugur Demir and Gozde Unal",

note = "Publisher Copyright: {\textcopyright} 2017 IEEE.; 25th Signal Processing and Communications Applications Conference, SIU 2017 ; Conference date: 15-05-2017 Through 18-05-2017",

year = "2017",

month = jun,

day = "27",

doi = "10.1109/SIU.2017.7960317",

language = "T{\"u}rk{\c c}e",

series = "2017 25th Signal Processing and Communications Applications Conference, SIU 2017",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "2017 25th Signal Processing and Communications Applications Conference, SIU 2017",

address = "United States",

}

Demir, U & Unal, G 2017, Danişman Aǧ Kullanarak Derin Otomatik Kodlayicilar ile İçboyama. in 2017 25th Signal Processing and Communications Applications Conference, SIU 2017., 7960317, 2017 25th Signal Processing and Communications Applications Conference, SIU 2017, Institute of Electrical and Electronics Engineers Inc., 25th Signal Processing and Communications Applications Conference, SIU 2017, Antalya, Turkey, 15/05/17. https://doi.org/10.1109/SIU.2017.7960317

Danişman Aǧ Kullanarak Derin Otomatik Kodlayicilar ile İçboyama. / Demir, Ugur; Unal, Gozde.
2017 25th Signal Processing and Communications Applications Conference, SIU 2017. Institute of Electrical and Electronics Engineers Inc., 2017. 7960317 (2017 25th Signal Processing and Communications Applications Conference, SIU 2017).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Danişman Aǧ Kullanarak Derin Otomatik Kodlayicilar ile İçboyama

AU - Demir, Ugur

AU - Unal, Gozde

PY - 2017/6/27

Y1 - 2017/6/27

N2 - In this study, we tried to find a solution for inpainting problem using deep convolutional autoencoders. A new training approach has been proposed as an alternative to the Generative Adversarial Networks. The neural network that designed for inpainting takes an image, which the certain part of its center is extracted, as an input then it attempts to fill the blank region. During the training phase, a distinct deep convolutional neural network is used and it is called Advisor Network. We show that the features extracted from intermediate layers of the Advisor Network, which is trained on a different dataset for classification, improves the performance of the autoencoder.

AB - In this study, we tried to find a solution for inpainting problem using deep convolutional autoencoders. A new training approach has been proposed as an alternative to the Generative Adversarial Networks. The neural network that designed for inpainting takes an image, which the certain part of its center is extracted, as an input then it attempts to fill the blank region. During the training phase, a distinct deep convolutional neural network is used and it is called Advisor Network. We show that the features extracted from intermediate layers of the Advisor Network, which is trained on a different dataset for classification, improves the performance of the autoencoder.

KW - auto-encoder

KW - convolutional neural network

KW - deep learning

KW - inpainting

UR - http://www.scopus.com/inward/record.url?scp=85026315710&partnerID=8YFLogxK

U2 - 10.1109/SIU.2017.7960317

DO - 10.1109/SIU.2017.7960317

M3 - Konferans katkısı

AN - SCOPUS:85026315710

T3 - 2017 25th Signal Processing and Communications Applications Conference, SIU 2017

BT - 2017 25th Signal Processing and Communications Applications Conference, SIU 2017

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 25th Signal Processing and Communications Applications Conference, SIU 2017

Y2 - 15 May 2017 through 18 May 2017

ER -

Danişman Aǧ Kullanarak Derin Otomatik Kodlayicilar ile İçboyama

Abstract

Publication series

Conference

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this