Learning Good Regions to Deblur Images

Research output: Contribution to journalArticle

14 Citations (Scopus)

Abstract

The goal of single image deblurring is to recover both a latent clear image and an underlying blur kernel from one input blurred image. Recent methods focus on exploiting natural image priors or additional image observations for deblurring, but pay less attention to the influence of image structure on estimating blur kernels. What is the useful image structure and how can one select good regions for deblurring? We formulate the problem of learning good regions for deblurring within the conditional random field framework. To better compare blur kernels, we develop an effective similarity metric for labeling training samples. The learned model is able to predict good regions from an input blurred image for deblurring without user guidance. Qualitative and quantitative evaluations demonstrate that good regions can be selected by the proposed algorithms for effective single image deblurring.

Original languageEnglish
Pages (from-to)345-362
Number of pages18
JournalInternational Journal of Computer Vision
Volume115
Issue number3
DOIs
Publication statusPublished - 2015 Dec 1

Fingerprint

Labeling

All Science Journal Classification (ASJC) codes

  • Software
  • Computer Vision and Pattern Recognition
  • Artificial Intelligence

Cite this

@article{b63caf99414a4f27b6b3794b12c3abc4,
title = "Learning Good Regions to Deblur Images",
abstract = "The goal of single image deblurring is to recover both a latent clear image and an underlying blur kernel from one input blurred image. Recent methods focus on exploiting natural image priors or additional image observations for deblurring, but pay less attention to the influence of image structure on estimating blur kernels. What is the useful image structure and how can one select good regions for deblurring? We formulate the problem of learning good regions for deblurring within the conditional random field framework. To better compare blur kernels, we develop an effective similarity metric for labeling training samples. The learned model is able to predict good regions from an input blurred image for deblurring without user guidance. Qualitative and quantitative evaluations demonstrate that good regions can be selected by the proposed algorithms for effective single image deblurring.",
author = "Zhe Hu and Yang, {Ming Hsuan}",
year = "2015",
month = "12",
day = "1",
doi = "10.1007/s11263-015-0821-1",
language = "English",
volume = "115",
pages = "345--362",
journal = "International Journal of Computer Vision",
issn = "0920-5691",
publisher = "Springer Netherlands",
number = "3",

}

Learning Good Regions to Deblur Images. / Hu, Zhe; Yang, Ming Hsuan.

In: International Journal of Computer Vision, Vol. 115, No. 3, 01.12.2015, p. 345-362.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Learning Good Regions to Deblur Images

AU - Hu, Zhe

AU - Yang, Ming Hsuan

PY - 2015/12/1

Y1 - 2015/12/1

N2 - The goal of single image deblurring is to recover both a latent clear image and an underlying blur kernel from one input blurred image. Recent methods focus on exploiting natural image priors or additional image observations for deblurring, but pay less attention to the influence of image structure on estimating blur kernels. What is the useful image structure and how can one select good regions for deblurring? We formulate the problem of learning good regions for deblurring within the conditional random field framework. To better compare blur kernels, we develop an effective similarity metric for labeling training samples. The learned model is able to predict good regions from an input blurred image for deblurring without user guidance. Qualitative and quantitative evaluations demonstrate that good regions can be selected by the proposed algorithms for effective single image deblurring.

AB - The goal of single image deblurring is to recover both a latent clear image and an underlying blur kernel from one input blurred image. Recent methods focus on exploiting natural image priors or additional image observations for deblurring, but pay less attention to the influence of image structure on estimating blur kernels. What is the useful image structure and how can one select good regions for deblurring? We formulate the problem of learning good regions for deblurring within the conditional random field framework. To better compare blur kernels, we develop an effective similarity metric for labeling training samples. The learned model is able to predict good regions from an input blurred image for deblurring without user guidance. Qualitative and quantitative evaluations demonstrate that good regions can be selected by the proposed algorithms for effective single image deblurring.

UR - http://www.scopus.com/inward/record.url?scp=84947024505&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84947024505&partnerID=8YFLogxK

U2 - 10.1007/s11263-015-0821-1

DO - 10.1007/s11263-015-0821-1

M3 - Article

AN - SCOPUS:84947024505

VL - 115

SP - 345

EP - 362

JO - International Journal of Computer Vision

JF - International Journal of Computer Vision

SN - 0920-5691

IS - 3

ER -