Model-based neural evaluation and iterative gradient optimization in image restoration and statistical filtering (original) (raw)

Abstract

An optimal model-based neural evaluation algorithm and an iterative gradient optimization algorithm used in image restoration and statistical filtering are presented. The relationship between the two algorithms is studied. We show that under the symmetric positive-definite condition, a condition easily satisfied in restoration and filtering, intra-pixelsequentialprocessing (IPSP) of model-based neuron evaluation is equivalent to the iterative gradient optimization algorithm. We also show that although both methods provide feasible solutions to fast spatial domain implementation of restoration and filtering techniques, the iterative gradient algorithm is in fact more efficient than the IPSP neuron evaluation method. Visual examples are provided to compare the performance of the two approaches.

Loading...

Loading Preview

Sorry, preview is currently unavailable. You can download the paper by clicking the button above.

References (20)

  1. H. c. Andrews and B. R. Hunt, Digital Image Restoration, Prentice- Hall, Englewood cliffs, NJ (1977).
  2. W. K. Pratt, Digital Signal Processing, 2nd ed., John Wiley & Sons, New York (1991).
  3. A. Katsaggelos, J. Biemond, R. Schafer, and R. Mersereau, ' 'A regu- larized iterative image restoration algorithm,' ' IEEE Trans. Signal Proc. SP-39(4), 914-929 (1991).
  4. J. W. Woods and V. K. Ingle, "Kalman filtering in two dimensions: further results," IEEE Trans. Acoust. Speech Sig. Proc. ASSP-29, 188-197 (1981).
  5. B. R. Suresh and B. A. Shenoi, ' 'New results in two dimensional Kalman filtering with application to image restoration,' ' IEEE Trans. Circuits Sys. CAS-28, 307-318 (1981).
  6. C. Mead, and M. A. Mahowald, "A silicon model for early visual processing," Neural Networks 1(1), 91-97 (1988).
  7. M. Bisio, D. D. Caviglia, G. Indiveri, L. Raffo, and S. P. Sabatini, ' 'Aneural model of cortical cells characterized by Gabor-like receptive fields-application to texture segmentation,' ' Proc. ICANN'92, pp. 917-920, Brighton, UK (1992).
  8. T. Wang, X. Zhuang, and X. Xing, ' 'Robust segmentation of noisy images using a neural network model,' ' Image Vision Computing 10(4), 233-240 (1992).
  9. J. K. Paik and A. K. Katsaggelos, "Edge detection using a neural network," Proc. ICASSP '90, pp. 2145-2148, Albuquerque (1990).
  10. 1 . S. W. Lu and A. Szeto, ''Hierarchical artificial neural networks for edge enhancement," Pattern Recog. 26(8), 1149-1163 (1993).
  11. L. Guan, "A neural network approach to adaptive image enhancement," Proc. SPIE 1658, 258-266, San Jose, CA (1992).
  12. Y. T. Zhou, R. Chellappa, A. Vaid, and B. K. Jenkins, "Image resto- ration using a neural network," IEEE Trans. ASSP ASSP-36(7), 1141-1 151 (1988).
  13. J. K. Paik and A. K. Katsaggelos, "Image restoration using a modified Hopfield network," IEEE Trans. Image Proc. 1(1), 49-63 (1992).
  14. 5. Rastogi and J. W. Woods, "Image restoration by parallel simulated annealing using compound Gauss-Markov models," Proc. ICCASP '91, pp. 2961-2964, Toronto, Canada (1991).
  15. J. Zhang, "The mean field theory in EM procedures for blind Markov random field image restoration," IEEE Trans. Image Proc. 2(1), 27-40 (1993).
  16. /Journalof Electronic Imaging/October 1995/ Vol. 4(4) Guan 17. L. Guan, ' 'Image restoration by a neural network with hierarchical cluster architecture," J. Electron. Imag. 3(2), 154-163 (1994).
  17. L. Guan, "Real-time image regularization: a neural computing per- spective to filtering, enhancement, and restoration,' ' Real-Time Im- aging: Theory, Techniques and Applications, P. Laplante and A. Stoy- enko, Eds., IEEE and IEEE/CS Press (Oct. 1995).
  18. J. P. Sutton, J. S. Beis, and L. E. H. Trainor, ' 'Hierarchical model of memory and memory loss,' ' J. Phys. A: Math. Gen. 21, 4443-4454 (1988).
  19. J. J. Hopfield and D. W. Tank, "Neural computation of decisions in optimization problems," Biol. Cybern. 52, 141-152 (1985).
  20. L. Guan and X. Zhou, "Real-time image filtering: from optimal neural evaluation to vector quadratic programming,' ' in Proc. IEEE Int. Conf on Systems, Men, and Cybernetics, pp. 694-699, San Antonio, TX (1994). Ling Guan received the BSc from Tianjin University, China, the MASc from Univer- sity of Waterloo, Canada, and a PhD from University of British Columbia, Canada. He is currently a lecturer and the director of the machine vision and real-time image processing group in the Department of Electrical Engineering at the University of Sydney, Australia. He previously worked as a senior research engineer at Array Systems Computing, Inc., Canada. He was a British Telecom Fellowship recipient in 1994. His research interests include image and signal processing, real-time machine vision, neural networks, multiresolution signal processing, and fuzzy logics.