TY - GEN
T1 - Debluring low-resolution images
AU - Pan, Jinshan
AU - Hu, Zhe
AU - Su, Zhixun
AU - Yang, Ming Hsuan
PY - 2017/1/1
Y1 - 2017/1/1
N2 - 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.
AB - 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.
UR - http://www.scopus.com/inward/record.url?scp=85016170335&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85016170335&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-54407-6_8
DO - 10.1007/978-3-319-54407-6_8
M3 - Conference contribution
AN - SCOPUS:85016170335
SN - 9783319544069
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 111
EP - 127
BT - Computer Vision - ACCV 2016 Workshops - ACCV 2016 International Workshops, Revised Selected Papers
A2 - Lu, Jiwen
A2 - Ma, Kai-Kuang
A2 - Chen, Chu-Song
PB - Springer Verlag
T2 - 13th Asian Conference on Computer Vision, ACCV 2016
Y2 - 20 November 2016 through 24 November 2016
ER -