A Fast BPNN Based Image Deblurring Method

DOI

10.2991/iccsee.2013.782How to use a DOI?

Keywords

image deblurring, back propagation neural networks, blur function.

Abstract

This paper proposes a fast deblurring method based on back propagation neural networks (BPNN), in which the symmetry of blur function is utilized to reduce the size of network. In the image block used to be training vector, the pixels that have the same distance with centrral pixel are set the same weight in BPNN. This set of weight is realized by adding all the pxiels that have the same distance to be one input data of BPNN. Comparing with the classic BPNN based deblurring method in which the symmetry of blur function is not considered, the proposed method decreased the computation consumption largely. At the same time, the performance of proposed method is improved. Several experiments results testify the superiority of proposed deblurring method.

Copyright

© 2013, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Xiaozhi Ren
AU  - Binchao Bin
PY  - 2013/03
DA  - 2013/03
TI  - A Fast BPNN Based Image Deblurring Method
BT  - Proceedings of the 2nd International Conference on Computer Science and Electronics Engineering (ICCSEE 2013)
PB  - Atlantis Press
SP  - 2976
EP  - 2979
SN  - 1951-6851
UR  - https://doi.org/10.2991/iccsee.2013.782
DO  - 10.2991/iccsee.2013.782
ID  - Ren2013/03
ER  -