Efficient Pyramid-Structured Networks Using Channel Separating Gate

Yoon Jae Yeo; Min Cheol Sagong; Sung Jea Ko; Dong Pan Lim

doi:10.1109/GCCE53005.2021.9621933

Efficient Pyramid-Structured Networks Using Channel Separating Gate

Yoon Jae Yeo, Min Cheol Sagong, Sung Jea Ko, Dong Pan Lim

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

Abstract

Convolutional neural networks (CNNs) have been brought noticeable improvement on numerous image processing applications. In particular, spatial pyramid structures have been widely used, but the number of required parameters increases exponentially as the network deepens. Therefore, we propose a channel separating gate (CSG) that can determine the number of channels in convolutional layers according to the spatial level. For a given spatial level, the CSG separates the channels of feature maps for use at the smaller level. Consequently, the standard non- pooling model can be altered to an efficient pyramid structure by performing the proposed method repeatedly. For single image super-resolution (SR) and deblurring, we demonstrate that our method improves the baseline network in terms of peak signal- to-noise ratio (PSNR) and floating-point operations (FLOPs).

Original language	English
Title of host publication	2021 IEEE 10th Global Conference on Consumer Electronics, GCCE 2021
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	986-987
Number of pages	2
ISBN (Electronic)	9781665436762
DOIs	https://doi.org/10.1109/GCCE53005.2021.9621933
Publication status	Published - 2021
Event	10th IEEE Global Conference on Consumer Electronics, GCCE 2021 - Kyoto, Japan Duration: 2021 Oct 12 → 2021 Oct 15

Publication series

Name	2021 IEEE 10th Global Conference on Consumer Electronics, GCCE 2021

Conference

Conference	10th IEEE Global Conference on Consumer Electronics, GCCE 2021
Country/Territory	Japan
City	Kyoto
Period	21/10/12 → 21/10/15

Bibliographical note

Publisher Copyright:
© 2021 IEEE.

Keywords

convolutional neural network (CNN)
deblurring
pyramid structure
super-resolution (SR)

ASJC Scopus subject areas

Computer Science Applications
Signal Processing
Biomedical Engineering
Electrical and Electronic Engineering
Media Technology
Instrumentation

Access to Document

10.1109/GCCE53005.2021.9621933

Cite this

Yeo, Y. J., Sagong, M. C., Ko, S. J., & Lim, D. P. (2021). Efficient Pyramid-Structured Networks Using Channel Separating Gate. In 2021 IEEE 10th Global Conference on Consumer Electronics, GCCE 2021 (pp. 986-987). (2021 IEEE 10th Global Conference on Consumer Electronics, GCCE 2021). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/GCCE53005.2021.9621933

Efficient Pyramid-Structured Networks Using Channel Separating Gate. / Yeo, Yoon Jae; Sagong, Min Cheol; Ko, Sung Jea et al.
2021 IEEE 10th Global Conference on Consumer Electronics, GCCE 2021. Institute of Electrical and Electronics Engineers Inc., 2021. p. 986-987 (2021 IEEE 10th Global Conference on Consumer Electronics, GCCE 2021).

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

Yeo, YJ, Sagong, MC, Ko, SJ & Lim, DP 2021, Efficient Pyramid-Structured Networks Using Channel Separating Gate. in 2021 IEEE 10th Global Conference on Consumer Electronics, GCCE 2021. 2021 IEEE 10th Global Conference on Consumer Electronics, GCCE 2021, Institute of Electrical and Electronics Engineers Inc., pp. 986-987, 10th IEEE Global Conference on Consumer Electronics, GCCE 2021, Kyoto, Japan, 21/10/12. https://doi.org/10.1109/GCCE53005.2021.9621933

@inproceedings{cc9a019eab464d6abe4cb827a36365e5,

title = "Efficient Pyramid-Structured Networks Using Channel Separating Gate",

abstract = "Convolutional neural networks (CNNs) have been brought noticeable improvement on numerous image processing applications. In particular, spatial pyramid structures have been widely used, but the number of required parameters increases exponentially as the network deepens. Therefore, we propose a channel separating gate (CSG) that can determine the number of channels in convolutional layers according to the spatial level. For a given spatial level, the CSG separates the channels of feature maps for use at the smaller level. Consequently, the standard non- pooling model can be altered to an efficient pyramid structure by performing the proposed method repeatedly. For single image super-resolution (SR) and deblurring, we demonstrate that our method improves the baseline network in terms of peak signal- to-noise ratio (PSNR) and floating-point operations (FLOPs).",

keywords = "convolutional neural network (CNN), deblurring, pyramid structure, super-resolution (SR)",

author = "Yeo, {Yoon Jae} and Sagong, {Min Cheol} and Ko, {Sung Jea} and Lim, {Dong Pan}",

note = "Publisher Copyright: {\textcopyright} 2021 IEEE.; 10th IEEE Global Conference on Consumer Electronics, GCCE 2021 ; Conference date: 12-10-2021 Through 15-10-2021",

year = "2021",

doi = "10.1109/GCCE53005.2021.9621933",

language = "English",

series = "2021 IEEE 10th Global Conference on Consumer Electronics, GCCE 2021",

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

pages = "986--987",

booktitle = "2021 IEEE 10th Global Conference on Consumer Electronics, GCCE 2021",

}

TY - GEN

T1 - Efficient Pyramid-Structured Networks Using Channel Separating Gate

AU - Yeo, Yoon Jae

AU - Sagong, Min Cheol

AU - Ko, Sung Jea

AU - Lim, Dong Pan

PY - 2021

Y1 - 2021

N2 - Convolutional neural networks (CNNs) have been brought noticeable improvement on numerous image processing applications. In particular, spatial pyramid structures have been widely used, but the number of required parameters increases exponentially as the network deepens. Therefore, we propose a channel separating gate (CSG) that can determine the number of channels in convolutional layers according to the spatial level. For a given spatial level, the CSG separates the channels of feature maps for use at the smaller level. Consequently, the standard non- pooling model can be altered to an efficient pyramid structure by performing the proposed method repeatedly. For single image super-resolution (SR) and deblurring, we demonstrate that our method improves the baseline network in terms of peak signal- to-noise ratio (PSNR) and floating-point operations (FLOPs).

AB - Convolutional neural networks (CNNs) have been brought noticeable improvement on numerous image processing applications. In particular, spatial pyramid structures have been widely used, but the number of required parameters increases exponentially as the network deepens. Therefore, we propose a channel separating gate (CSG) that can determine the number of channels in convolutional layers according to the spatial level. For a given spatial level, the CSG separates the channels of feature maps for use at the smaller level. Consequently, the standard non- pooling model can be altered to an efficient pyramid structure by performing the proposed method repeatedly. For single image super-resolution (SR) and deblurring, we demonstrate that our method improves the baseline network in terms of peak signal- to-noise ratio (PSNR) and floating-point operations (FLOPs).

KW - convolutional neural network (CNN)

KW - deblurring

KW - pyramid structure

KW - super-resolution (SR)

UR - http://www.scopus.com/inward/record.url?scp=85123476951&partnerID=8YFLogxK

U2 - 10.1109/GCCE53005.2021.9621933

DO - 10.1109/GCCE53005.2021.9621933

M3 - Conference contribution

AN - SCOPUS:85123476951

T3 - 2021 IEEE 10th Global Conference on Consumer Electronics, GCCE 2021

SP - 986

EP - 987

BT - 2021 IEEE 10th Global Conference on Consumer Electronics, GCCE 2021

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 10th IEEE Global Conference on Consumer Electronics, GCCE 2021

Y2 - 12 October 2021 through 15 October 2021

ER -

Efficient Pyramid-Structured Networks Using Channel Separating Gate

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this