Share Email Print
cover

Proceedings Paper

Filter designing for image processing based on multidimensional linear extrapolation
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

This article is proposed to develop a method of the synthesis of the prediction filter based on the multidimensional linear extrapolation to improve the accuracy of the pixel value prediction, as well as to assess the effectiveness of the proposed predictive filters depending on the prediction step and the dimension of training vectors to minimize the prediction error and the entropy of the differential signal. The article contain a solution of the special cases for a number of elements of the vector prediction, statistical dependence of the vector dimension and prediction filter order from the mean square error and the entropy of the differential signal are determined. Optimal prediction step is founded based on the experimental data and all the results are being analyzed.

Paper Details

Date Published: 4 April 2014
PDF: 6 pages
Proc. SPIE 9156, Optical Technologies for Telecommunications 2013, 91560I (4 April 2014); doi: 10.1117/12.2054871
Show Author Affiliations
Valery Kh. Bagmanov, Ufa State Aviation Technical Univ. (Russian Federation)
Airat R. Zainullin, Ufa State Aviation Technical Univ. (Russian Federation)
Ivan K. Meshkov, Ufa State Aviation Technical Univ. (Russian Federation)
Albert Kh. Sultanov, Ufa State Aviation Technical Univ. (Russian Federation)


Published in SPIE Proceedings Vol. 9156:
Optical Technologies for Telecommunications 2013
Vladimir A. Andreev; Vladimir A. Burdin; Albert H. Sultanov; Oleg G. Morozov, Editor(s)

© SPIE. Terms of Use
Back to Top