Learning fully convolutional networks for iterative non-blind deconvolution

Jiawei Zhang; Jinshan Pan; Wei Sheng Lai; Rynson W.H. Lau; Ming Hsuan Yang

doi:10.1109/CVPR.2017.737

Learning fully convolutional networks for iterative non-blind deconvolution

Jiawei Zhang, Jinshan Pan, Wei Sheng Lai, Rynson W.H. Lau, Ming Hsuan Yang

College of Computing

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

113 Citations (Scopus)

Abstract

In this paper, we propose a fully convolutional network for iterative non-blind deconvolution. We decompose the non-blind deconvolution problem into image denoising and image deconvolution. We train a FCNN to remove noise in the gradient domain and use the learned gradients to guide the image deconvolution step. In contrast to the existing deep neural network based methods, we iteratively deconvolve the blurred images in a multi-stage framework. The proposed method is able to learn an adaptive image prior, which keeps both local (details) and global (structures) information. Both quantitative and qualitative evaluations on the benchmark datasets demonstrate that the proposed method performs favorably against state-of-the-art algorithms in terms of quality and speed.

Original language	English
Title of host publication	Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	6969-6977
Number of pages	9
ISBN (Electronic)	9781538604571
DOIs	https://doi.org/10.1109/CVPR.2017.737
Publication status	Published - 2017 Nov 6
Event	30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017 - Honolulu, United States Duration: 2017 Jul 21 → 2017 Jul 26

Publication series

Name	Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017
Volume	2017-January

Other

Other	30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017
Country/Territory	United States
City	Honolulu
Period	17/7/21 → 17/7/26

Bibliographical note

Funding Information:
Acknowledgements. This work is supported in part by the SRG grant from City University of Hong Kong (No. 7004416), the National Natural Science Foundation of China (No. 61572099 and 61320106008), the NSF Career Grant 1149783 and gifts from Adobe and Nvidia.

Funding Information:
This work is supported in part by the SRG grant from City University of Hong Kong (No. 7004416), the National Natural Science Foundation of China (No. 61572099 and 61320106008), the NSF Career Grant 1149783 and gifts from Adobe and Nvidia.

Publisher Copyright:
© 2017 IEEE.

All Science Journal Classification (ASJC) codes

Signal Processing
Computer Vision and Pattern Recognition

Access to Document

10.1109/CVPR.2017.737

Cite this

Zhang, J., Pan, J., Lai, W. S., Lau, R. W. H., & Yang, M. H. (2017). Learning fully convolutional networks for iterative non-blind deconvolution. In Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017 (pp. 6969-6977). (Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017; Vol. 2017-January). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CVPR.2017.737

Zhang, Jiawei ; Pan, Jinshan ; Lai, Wei Sheng et al. / Learning fully convolutional networks for iterative non-blind deconvolution. Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 6969-6977 (Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017).

@inproceedings{d1c50a0ca4154ffe96afce18a48f949c,

title = "Learning fully convolutional networks for iterative non-blind deconvolution",

abstract = "In this paper, we propose a fully convolutional network for iterative non-blind deconvolution. We decompose the non-blind deconvolution problem into image denoising and image deconvolution. We train a FCNN to remove noise in the gradient domain and use the learned gradients to guide the image deconvolution step. In contrast to the existing deep neural network based methods, we iteratively deconvolve the blurred images in a multi-stage framework. The proposed method is able to learn an adaptive image prior, which keeps both local (details) and global (structures) information. Both quantitative and qualitative evaluations on the benchmark datasets demonstrate that the proposed method performs favorably against state-of-the-art algorithms in terms of quality and speed.",

author = "Jiawei Zhang and Jinshan Pan and Lai, {Wei Sheng} and Lau, {Rynson W.H.} and Yang, {Ming Hsuan}",

note = "Funding Information: Acknowledgements. This work is supported in part by the SRG grant from City University of Hong Kong (No. 7004416), the National Natural Science Foundation of China (No. 61572099 and 61320106008), the NSF Career Grant 1149783 and gifts from Adobe and Nvidia. Funding Information: This work is supported in part by the SRG grant from City University of Hong Kong (No. 7004416), the National Natural Science Foundation of China (No. 61572099 and 61320106008), the NSF Career Grant 1149783 and gifts from Adobe and Nvidia. Publisher Copyright: {\textcopyright} 2017 IEEE.; 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017 ; Conference date: 21-07-2017 Through 26-07-2017",

year = "2017",

month = nov,

day = "6",

doi = "10.1109/CVPR.2017.737",

language = "English",

series = "Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017",

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

pages = "6969--6977",

booktitle = "Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017",

address = "United States",

}

Zhang, J, Pan, J, Lai, WS, Lau, RWH & Yang, MH 2017, Learning fully convolutional networks for iterative non-blind deconvolution. in Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017. Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017, vol. 2017-January, Institute of Electrical and Electronics Engineers Inc., pp. 6969-6977, 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017, Honolulu, United States, 17/7/21. https://doi.org/10.1109/CVPR.2017.737

Learning fully convolutional networks for iterative non-blind deconvolution. / Zhang, Jiawei; Pan, Jinshan; Lai, Wei Sheng et al.
Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017. Institute of Electrical and Electronics Engineers Inc., 2017. p. 6969-6977 (Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017; Vol. 2017-January).

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

TY - GEN

T1 - Learning fully convolutional networks for iterative non-blind deconvolution

AU - Zhang, Jiawei

AU - Pan, Jinshan

AU - Lai, Wei Sheng

AU - Lau, Rynson W.H.

AU - Yang, Ming Hsuan

N1 - Funding Information: Acknowledgements. This work is supported in part by the SRG grant from City University of Hong Kong (No. 7004416), the National Natural Science Foundation of China (No. 61572099 and 61320106008), the NSF Career Grant 1149783 and gifts from Adobe and Nvidia. Funding Information: This work is supported in part by the SRG grant from City University of Hong Kong (No. 7004416), the National Natural Science Foundation of China (No. 61572099 and 61320106008), the NSF Career Grant 1149783 and gifts from Adobe and Nvidia. Publisher Copyright: © 2017 IEEE.

PY - 2017/11/6

Y1 - 2017/11/6

N2 - In this paper, we propose a fully convolutional network for iterative non-blind deconvolution. We decompose the non-blind deconvolution problem into image denoising and image deconvolution. We train a FCNN to remove noise in the gradient domain and use the learned gradients to guide the image deconvolution step. In contrast to the existing deep neural network based methods, we iteratively deconvolve the blurred images in a multi-stage framework. The proposed method is able to learn an adaptive image prior, which keeps both local (details) and global (structures) information. Both quantitative and qualitative evaluations on the benchmark datasets demonstrate that the proposed method performs favorably against state-of-the-art algorithms in terms of quality and speed.

AB - In this paper, we propose a fully convolutional network for iterative non-blind deconvolution. We decompose the non-blind deconvolution problem into image denoising and image deconvolution. We train a FCNN to remove noise in the gradient domain and use the learned gradients to guide the image deconvolution step. In contrast to the existing deep neural network based methods, we iteratively deconvolve the blurred images in a multi-stage framework. The proposed method is able to learn an adaptive image prior, which keeps both local (details) and global (structures) information. Both quantitative and qualitative evaluations on the benchmark datasets demonstrate that the proposed method performs favorably against state-of-the-art algorithms in terms of quality and speed.

UR - http://www.scopus.com/inward/record.url?scp=85041819876&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85041819876&partnerID=8YFLogxK

U2 - 10.1109/CVPR.2017.737

DO - 10.1109/CVPR.2017.737

M3 - Conference contribution

AN - SCOPUS:85041819876

T3 - Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017

SP - 6969

EP - 6977

BT - Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017

Y2 - 21 July 2017 through 26 July 2017

ER -

Zhang J, Pan J, Lai WS, Lau RWH, Yang MH. Learning fully convolutional networks for iterative non-blind deconvolution. In Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017. Institute of Electrical and Electronics Engineers Inc. 2017. p. 6969-6977. (Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017). doi: 10.1109/CVPR.2017.737

Learning fully convolutional networks for iterative non-blind deconvolution

Abstract

Publication series

Other

Bibliographical note

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this