Conference paper

SVOBODA Pavel, HRADIŠ Michal, MARŠÍK Lukáš and ZEMČÍK Pavel. CNN for license plate motion deblurring. In: IEEE International Conference on Image Processing (ICIP). Phoenix: IEEE Signal Processing Society, 2016, pp. 1-4. ISBN 978-1-4673-9961-6. Available from:
Publication language:english
Original title:CNN for license plate motion deblurring
Title (cs):Konvoluční neuronové sítě pro restauraci registračních značek rozmazání pohybem
Proceedings:IEEE International Conference on Image Processing (ICIP)
Conference:International Conference on Image Processing (ICIP) 2016
Place:Phoenix, US
Publisher:IEEE Signal Processing Society
Convolutional neural network, Motion blur, Image reconstruction, Blind deconvolution, License plate
In this work we explore the previously proposed approach of direct blind deconvolution and denoising with convolutional neural networks (CNN) in a situation where the blur kernels are partially constrained. We focus on blurred images from a real-life traffic surveillance system, on which we, for the first time, demonstrate that neural networks trained on artificial data provide superior reconstruction quality on real images compared to traditional blind deconvolution methods. The training data is easy to obtain by blurring sharp photos from a target system with a very rough approximation of the expected blur kernels, thereby allowing custom CNNs to be trained for a specific application (image content and blur range). Additionally, we evaluate the behavior and limits of the CNNs with respect to blur direction range and length.
   author = {Pavel Svoboda and Michal Hradi{\v{s}} and
	Luk{\'{a}}{\v{s}} Mar{\v{s}}{\'{i}}k and Pavel
   title = {CNN for license plate motion deblurring},
   pages = {1--4},
   booktitle = {IEEE International Conference on Image Processing (ICIP)},
   year = 2016,
   location = {Phoenix, US},
   publisher = {IEEE Signal Processing Society},
   ISBN = {978-1-4673-9961-6},
   doi = {10.1109/ICIP.2016.7533077},
   language = {english},
   url = {}

Your IPv4 address:
Switch to https