Title
Blur detection via deep pyramid network with recurrent distinction enhanced modules
Abstract
•We present a new deep pyramid network to solve the blur detection problem.•The DEBs keep a desirable identification between the blur and clear regions.•We develop a new boundary penalty term to refine the boundaries of our results.•A new blur detection dataset, named SZU-BD, has been established.
Year
DOI
Venue
2020
10.1016/j.neucom.2020.06.068
Neurocomputing
Keywords
DocType
Volume
Blur detection,Feature pyramid network,Distinction enhanced block,Boundary-aware penalty,Deep learning
Journal
414
ISSN
Citations 
PageRank 
0925-2312
0
0.34
References 
Authors
0
4
Name
Order
Citations
PageRank
Xiaoli Sun1265.49
Xiujun Zhang215918.75
Mingqing Xiao342.79
Chen Xu426929.36