Motion Blur Kernel Estimation via Deep Learning

Xiangyu Xu, Jinshan Pan, Yu Jin Zhang, Ming Hsuan Yang

Research output: Contribution to journalArticle

23 Citations (Scopus)

Abstract

The success of the state-of-the-art deblurring methods mainly depends on the restoration of sharp edges in a coarse-to-fine kernel estimation process. In this paper, we propose to learn a deep convolutional neural network for extracting sharp edges from blurred images. Motivated by the success of the existing filtering-based deblurring methods, the proposed model consists of two stages: suppressing extraneous details and enhancing sharp edges. We show that the two-stage model simplifies the learning process and effectively restores sharp edges. Facilitated by the learned sharp edges, the proposed deblurring algorithm does not require any coarse-to-fine strategy or edge selection, thereby significantly simplifying kernel estimation and reducing computation load. Extensive experimental results on challenging blurry images demonstrate that the proposed algorithm performs favorably against the state-of-the-art methods on both synthetic and real-world images in terms of visual quality and run-time.

Original languageEnglish
Article number8039224
Pages (from-to)194-205
Number of pages12
JournalIEEE Transactions on Image Processing
Volume27
Issue number1
DOIs
Publication statusPublished - 2018 Jan

Fingerprint

Restoration
Neural networks
Deep learning

All Science Journal Classification (ASJC) codes

  • Software
  • Computer Graphics and Computer-Aided Design

Cite this

Xu, Xiangyu ; Pan, Jinshan ; Zhang, Yu Jin ; Yang, Ming Hsuan. / Motion Blur Kernel Estimation via Deep Learning. In: IEEE Transactions on Image Processing. 2018 ; Vol. 27, No. 1. pp. 194-205.
@article{25d2c11797704b95b0273b0cd4ed77e2,
title = "Motion Blur Kernel Estimation via Deep Learning",
abstract = "The success of the state-of-the-art deblurring methods mainly depends on the restoration of sharp edges in a coarse-to-fine kernel estimation process. In this paper, we propose to learn a deep convolutional neural network for extracting sharp edges from blurred images. Motivated by the success of the existing filtering-based deblurring methods, the proposed model consists of two stages: suppressing extraneous details and enhancing sharp edges. We show that the two-stage model simplifies the learning process and effectively restores sharp edges. Facilitated by the learned sharp edges, the proposed deblurring algorithm does not require any coarse-to-fine strategy or edge selection, thereby significantly simplifying kernel estimation and reducing computation load. Extensive experimental results on challenging blurry images demonstrate that the proposed algorithm performs favorably against the state-of-the-art methods on both synthetic and real-world images in terms of visual quality and run-time.",
author = "Xiangyu Xu and Jinshan Pan and Zhang, {Yu Jin} and Yang, {Ming Hsuan}",
year = "2018",
month = "1",
doi = "10.1109/TIP.2017.2753658",
language = "English",
volume = "27",
pages = "194--205",
journal = "IEEE Transactions on Image Processing",
issn = "1057-7149",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "1",

}

Motion Blur Kernel Estimation via Deep Learning. / Xu, Xiangyu; Pan, Jinshan; Zhang, Yu Jin; Yang, Ming Hsuan.

In: IEEE Transactions on Image Processing, Vol. 27, No. 1, 8039224, 01.2018, p. 194-205.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Motion Blur Kernel Estimation via Deep Learning

AU - Xu, Xiangyu

AU - Pan, Jinshan

AU - Zhang, Yu Jin

AU - Yang, Ming Hsuan

PY - 2018/1

Y1 - 2018/1

N2 - The success of the state-of-the-art deblurring methods mainly depends on the restoration of sharp edges in a coarse-to-fine kernel estimation process. In this paper, we propose to learn a deep convolutional neural network for extracting sharp edges from blurred images. Motivated by the success of the existing filtering-based deblurring methods, the proposed model consists of two stages: suppressing extraneous details and enhancing sharp edges. We show that the two-stage model simplifies the learning process and effectively restores sharp edges. Facilitated by the learned sharp edges, the proposed deblurring algorithm does not require any coarse-to-fine strategy or edge selection, thereby significantly simplifying kernel estimation and reducing computation load. Extensive experimental results on challenging blurry images demonstrate that the proposed algorithm performs favorably against the state-of-the-art methods on both synthetic and real-world images in terms of visual quality and run-time.

AB - The success of the state-of-the-art deblurring methods mainly depends on the restoration of sharp edges in a coarse-to-fine kernel estimation process. In this paper, we propose to learn a deep convolutional neural network for extracting sharp edges from blurred images. Motivated by the success of the existing filtering-based deblurring methods, the proposed model consists of two stages: suppressing extraneous details and enhancing sharp edges. We show that the two-stage model simplifies the learning process and effectively restores sharp edges. Facilitated by the learned sharp edges, the proposed deblurring algorithm does not require any coarse-to-fine strategy or edge selection, thereby significantly simplifying kernel estimation and reducing computation load. Extensive experimental results on challenging blurry images demonstrate that the proposed algorithm performs favorably against the state-of-the-art methods on both synthetic and real-world images in terms of visual quality and run-time.

UR - http://www.scopus.com/inward/record.url?scp=85030645947&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85030645947&partnerID=8YFLogxK

U2 - 10.1109/TIP.2017.2753658

DO - 10.1109/TIP.2017.2753658

M3 - Article

C2 - 28922121

AN - SCOPUS:85030645947

VL - 27

SP - 194

EP - 205

JO - IEEE Transactions on Image Processing

JF - IEEE Transactions on Image Processing

SN - 1057-7149

IS - 1

M1 - 8039224

ER -