27 January 2009 A novel edge-parameter analysis approach of blur identification for image de-blurring application
Author Affiliations +
Abstract
In this paper, a novel edge-parameter analysis method of the blu' identification based on the single-threshold Pulse Coupled Neural Networks (PCNN) model is proposed for image de-blurring application. It suits to the identification of the horizontal linear motion blur. This new identification method not only improves on the traditional PCNN, but also uses the normalized local entropy. On the one hand, the new method uses the local entropy which is normalized between 0 and 255. On the other hand, a new model called the single-threshold PCNN is proposed in this article. Comparing with the traditional PCNN, the improved one calculates faster, and it is more sensitive to the image edges. The experimental results which are obtained from the different images and the same image with the different resolution show that the new algorithm is very effective and the curve is the very steady graph. The identification precision is about 4 to 30 pixels.
© (2009) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Jinping He, Jinping He, Kun Gao, Kun Gao, Guoqiang Ni, Guoqiang Ni, } "A novel edge-parameter analysis approach of blur identification for image de-blurring application", Proc. SPIE 7156, 2008 International Conference on Optical Instruments and Technology: Optical Systems and Optoelectronic Instruments, 71561V (27 January 2009); doi: 10.1117/12.806711; https://doi.org/10.1117/12.806711
PROCEEDINGS
8 PAGES


SHARE
RELATED CONTENT

Neural network modelling by rank configurations
Proceedings of SPIE (September 30 2018)
Image Algebra And Its Relationship To Neural Networks
Proceedings of SPIE (August 29 1989)
Neural network-based multiscale image restoration approach
Proceedings of SPIE (February 26 2007)
Motion blur removal based on restoration error analysis
Proceedings of SPIE (February 28 2005)
Theory of morphological neural networks
Proceedings of SPIE (June 30 1990)
Affine morphology and affine signal models
Proceedings of SPIE (October 31 1990)

Back to Top