Debluring low-resolution images

Jinshan Pan, Zhe Hu, Zhixun Su, Ming Hsuan Yang

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

The recent years have witnessed significant advances in image deblurring. In general, the success of deblurring methods depends heavily on extraction of salient structures from a blurry image for kernel estimation. Most deblurring methods often operate on high-resolution images where contours or edges can be extracted for kernel estimation. However, recovering reliable structures from low-resolution images becomes extremely challenging. In this paper, we propose a spatially variant deblurring algorithm for low-resolution images based on the exemplars. To exploit the exemplar information, we develop a super-resolution guided method to help the restoration of reliable image structures which can be used for kernel estimation. Experimental evaluations against the state-of-the-art methods show that the proposed algorithm performs favorably in deblurring low-resolution images. Furthermore, we show that the SR results obtained as byproducts in our method are comparable compared to other blind SR methods.

Original languageEnglish
Title of host publicationComputer Vision - ACCV 2016 Workshops - ACCV 2016 International Workshops, Revised Selected Papers
EditorsJiwen Lu, Kai-Kuang Ma, Chu-Song Chen
PublisherSpringer Verlag
Pages111-127
Number of pages17
ISBN (Print)9783319544069
DOIs
Publication statusPublished - 2017
Event13th Asian Conference on Computer Vision, ACCV 2016 - Taipei, Taiwan, Province of China
Duration: 2016 Nov 202016 Nov 24

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10116 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other13th Asian Conference on Computer Vision, ACCV 2016
Country/TerritoryTaiwan, Province of China
City Taipei
Period16/11/2016/11/24

Bibliographical note

Funding Information:
This work has been supported in part by NSF CAREER (No. 1149783), NSF IIS (No. 1152576), NSFC (No. 61572099 and 61320106008) and a gift from Adobe.

Publisher Copyright:
© Springer International Publishing AG 2017.

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint

Dive into the research topics of 'Debluring low-resolution images'. Together they form a unique fingerprint.

Cite this