(26) * << * >> * Russian * English * Content * All Issues

Comparative analysis of function approximation methods in image processing tasks

V.V. Sergeev1, V.N. Kopenkov2, A.V. Chernov2
1Image Processing Systems Institute of RAS 

2Samara State Aerospace University 

 PDF, 114 kB

Pages: 119-122.

Abstract:
The paper considers various nonlinear methods of multivariate regression approximation (neural networks, linear parameter functions, hierarchical approximation) as applied to image filtering problems based on a priori information in the form of a pair of images (“ideal” + “distorted”). The considered approximation methods are compared in terms of efficiency.

Keywords:
image processing, nonlinear method, neural network, linear parameter function, hierarchical approximation.

Citation:
Sergeev VV, Kopenkov VN, Chernov AV. Comparative analysis of function approximation methods in image processing tasks. Computer Optics 2004; 26: 119-122.

References:
  1. Kalitkin NN. Numerical methods; Moscow: Nauka Publisher; 1978: 27-69. 
  2. Korotky S. Neural networks: back propagation algorithm http://lii.newmail.ru/NN/KOROTKY/N2/kor_nn2.htm. 
  3. Soifer VA, ed. Computer image processing methods [In Russian]. Moscow: "Fizmatlit" Publisher; 2001: 527-598. 
  4. Krose B, van der Smagt P. An introduction to Neural network. 8th ed. Amsterdam: The University of Amsterdam; 1996: 15-55. 
  5. Sergeyev VV, Chernov AV. Image reconstruction methods based on the principles of pattern recognition theory. Pattern Recognition and Image Analysis 1997; 7(4): 474-479.

© 2009, IPSI RAS
151, Molodogvardeiskaya str., Samara, 443001, Russia; E-mail: ko@smr.ru ; Tel: +7 (846) 242-41-24 (Executive secretary), +7 (846) 332-56-22 (Issuing editor), Fax: +7 (846) 332-56-20