Joint super-resolution and compression artifact reduction based on dual-learning

Oh Young Lee; Jae Won Lee; Dae Yeol Lee; Jong-Ok Kim

doi:10.1109/VCIP.2016.7805543

Joint super-resolution and compression artifact reduction based on dual-learning

Oh Young Lee, Jae Won Lee, Dae Yeol Lee, Jong-Ok Kim

School of Electrical Engineering

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

3 Citations (Scopus)

Abstract

We propose a novel integrated framework to combine the self-learning super-resolution (SR) with dual-learning noise-reduction (NR) for compressed images. Contrary to existing learning based denoising approach, dual-learning based joint SR and NR is proposed by adding a denoised training set. It makes the proposed framework more suitable for highly compressed noise by referring to closer patch in a training set. Also, it is robust for SR artifacts since the joint framework is designed in such a way that one could learn a process to simultaneously perform NR and SR. Experimental results show that the proposed joint SR and NR framework can achieve higher objective and subjective qualities, compared with individual processing of NR and SR.

Original language	English
Title of host publication	VCIP 2016 - 30th Anniversary of Visual Communication and Image Processing
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9781509053162
DOIs	https://doi.org/10.1109/VCIP.2016.7805543
Publication status	Published - 2017 Jan 4
Event	2016 IEEE Visual Communication and Image Processing, VCIP 2016 - Chengdu, China Duration: 2016 Nov 27 → 2016 Nov 30

Other

Other	2016 IEEE Visual Communication and Image Processing, VCIP 2016
Country	China
City	Chengdu
Period	16/11/27 → 16/11/30

Keywords

compression artifact reduction
Dual-learning
image super resolution
self-learning

ASJC Scopus subject areas

Computer Networks and Communications
Computer Vision and Pattern Recognition
Signal Processing

Access to Document

10.1109/VCIP.2016.7805543

Fingerprint Dive into the research topics of 'Joint super-resolution and compression artifact reduction based on dual-learning'. Together they form a unique fingerprint.

Cite this

Lee, O. Y., Lee, J. W., Lee, D. Y., & Kim, J-O. (2017). Joint super-resolution and compression artifact reduction based on dual-learning. In VCIP 2016 - 30th Anniversary of Visual Communication and Image Processing [7805543] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/VCIP.2016.7805543

Joint super-resolution and compression artifact reduction based on dual-learning. / Lee, Oh Young; Lee, Jae Won; Lee, Dae Yeol; Kim, Jong-Ok.

VCIP 2016 - 30th Anniversary of Visual Communication and Image Processing. Institute of Electrical and Electronics Engineers Inc., 2017. 7805543.

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

Lee, OY, Lee, JW, Lee, DY & Kim, J-O 2017, Joint super-resolution and compression artifact reduction based on dual-learning. in VCIP 2016 - 30th Anniversary of Visual Communication and Image Processing., 7805543, Institute of Electrical and Electronics Engineers Inc., 2016 IEEE Visual Communication and Image Processing, VCIP 2016, Chengdu, China, 16/11/27. https://doi.org/10.1109/VCIP.2016.7805543

@inproceedings{7a335a075dce44b8b3502f456d1dfa4f,

title = "Joint super-resolution and compression artifact reduction based on dual-learning",

abstract = "We propose a novel integrated framework to combine the self-learning super-resolution (SR) with dual-learning noise-reduction (NR) for compressed images. Contrary to existing learning based denoising approach, dual-learning based joint SR and NR is proposed by adding a denoised training set. It makes the proposed framework more suitable for highly compressed noise by referring to closer patch in a training set. Also, it is robust for SR artifacts since the joint framework is designed in such a way that one could learn a process to simultaneously perform NR and SR. Experimental results show that the proposed joint SR and NR framework can achieve higher objective and subjective qualities, compared with individual processing of NR and SR.",

keywords = "compression artifact reduction, Dual-learning, image super resolution, self-learning",

author = "Lee, {Oh Young} and Lee, {Jae Won} and Lee, {Dae Yeol} and Jong-Ok Kim",

year = "2017",

month = jan,

day = "4",

doi = "10.1109/VCIP.2016.7805543",

language = "English",

booktitle = "VCIP 2016 - 30th Anniversary of Visual Communication and Image Processing",

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

note = "2016 IEEE Visual Communication and Image Processing, VCIP 2016 ; Conference date: 27-11-2016 Through 30-11-2016",

}

TY - GEN

T1 - Joint super-resolution and compression artifact reduction based on dual-learning

AU - Lee, Oh Young

AU - Lee, Jae Won

AU - Lee, Dae Yeol

AU - Kim, Jong-Ok

PY - 2017/1/4

Y1 - 2017/1/4

N2 - We propose a novel integrated framework to combine the self-learning super-resolution (SR) with dual-learning noise-reduction (NR) for compressed images. Contrary to existing learning based denoising approach, dual-learning based joint SR and NR is proposed by adding a denoised training set. It makes the proposed framework more suitable for highly compressed noise by referring to closer patch in a training set. Also, it is robust for SR artifacts since the joint framework is designed in such a way that one could learn a process to simultaneously perform NR and SR. Experimental results show that the proposed joint SR and NR framework can achieve higher objective and subjective qualities, compared with individual processing of NR and SR.

AB - We propose a novel integrated framework to combine the self-learning super-resolution (SR) with dual-learning noise-reduction (NR) for compressed images. Contrary to existing learning based denoising approach, dual-learning based joint SR and NR is proposed by adding a denoised training set. It makes the proposed framework more suitable for highly compressed noise by referring to closer patch in a training set. Also, it is robust for SR artifacts since the joint framework is designed in such a way that one could learn a process to simultaneously perform NR and SR. Experimental results show that the proposed joint SR and NR framework can achieve higher objective and subjective qualities, compared with individual processing of NR and SR.

KW - compression artifact reduction

KW - Dual-learning

KW - image super resolution

KW - self-learning

UR - http://www.scopus.com/inward/record.url?scp=85011072032&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85011072032&partnerID=8YFLogxK

U2 - 10.1109/VCIP.2016.7805543

DO - 10.1109/VCIP.2016.7805543

M3 - Conference contribution

AN - SCOPUS:85011072032

BT - VCIP 2016 - 30th Anniversary of Visual Communication and Image Processing

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2016 IEEE Visual Communication and Image Processing, VCIP 2016

Y2 - 27 November 2016 through 30 November 2016

ER -