Neural networks for linear inverse problems with incomplete data especially in applications to signal and image reconstruction

Abstract In this paper a generalized minimum norm principle and/or a maximum entropy principle are used as general criteria, e.g. for reconstructing images or signals from noisy and incomplete projection data. The linear inverse problem is reformulated as an optimization problem and solved by a unified analog neural network. Moreover, in order to reduce the network complexity a new approach is proposed which makes it possible to construct simply and effectively a neural network containing only one single artificial neuron with an on chip adaptive learning algorithm. The correctness and performance of the proposed neural network architectures are illustrated by extensive computer simulation experiments.

[1]  Gene H. Golub,et al.  Numerical Linear Algebra, Digital Signal Processing and Parallel Algorithms , 1991, NATO ASI Series.

[2]  Stefen Hui,et al.  Neural networks for constrained optimization problems , 1993, Int. J. Circuit Theory Appl..

[3]  J. Ramirez-Angulo,et al.  Architectures for analog VLSI implementation of neural networks for solving linear equations with inequality constraints , 1992, [Proceedings] 1992 IEEE International Symposium on Circuits and Systems.

[4]  D. Ingman,et al.  Maximum entropy signal reconstruction with neural networks , 1992, IEEE Trans. Neural Networks.

[5]  S. Levy,et al.  Reconstruction of a sparse spike train from a portion of its spectrum and application to high-resolution deconvolution , 1981 .

[6]  Andrzej Cichocki,et al.  Winner-take-all neural network and its application to minimax optimisation problems , 1991 .

[7]  Stanislaw Osowski,et al.  Neural networks in interpolation problems , 1993, Neurocomputing.

[8]  W. R. Madych.  On Underdetermined Systems , 1991 .

[9]  Achiya Dax.  The Convergence of Linear Stationary Iterative Processes for Solving Singular Unstructured Systems of Linear Equations , 1990, SIAM Rev..

[10]  Y. Censor,et al.  On iterative methods for linearly constrained entropy maximization , 1990 .

[11]  J. Hopfield.  Neurons withgraded response havecollective computational properties likethoseoftwo-state neurons , 1984 .

[12]  Satish S. Udpa,et al.  Projection iterative reconstruction technique and optoelectronic implementation , 1992, [Proceedings] 1992 IEEE International Symposium on Circuits and Systems.

[13]  Andrzej Cichocki,et al.  Neural networks for optimization and signal processing , 1993 .

[14]  John J. Hopfield,et al.  Simple 'neural' optimization networks: An A/D converter, signal decision circuit, and a linear programming circuit , 1986 .

[15]  C. Portmann,et al.  Practical design and analysis of a simple 'neural' optimization circuit , 1989 .

[16]  Richard J. Mammone.  Computational methods of signal recovery and recognition , 1992 .

[17]  A. Cichocki,et al.  Neural networks for solving systems of linear equations and related problems , 1992 .

[18]  B. K. Jenkins,et al.  Image restoration using a neural network , 1988, IEEE Trans. Acoust. Speech Signal Process..

[19]  Leon O. Chua,et al.  Neural networks for nonlinear programming , 1988 .

[20]  Andrzej Cichocki,et al.  Switched-capacitor neural networks for differential optimization , 1991, Int. J. Circuit Theory Appl..

[21]  J. P. Lasalle.  The stability and control of discrete processes , 1986 .

[22]  Andrzej Cichocki,et al.  Linear Least Least Simplified Neural Networks for Solving Squares Problems in Real Time Squares and Total , 1994 .

[23]  Janina Kowalski,et al.  Numerical analysis and mathematical modelling , 1990 .

[24]  Andrzej Cichocki,et al.  Switched-capacitor artificial neural networks for nonlinear optimization with constraints , 1990, IEEE International Symposium on Circuits and Systems.

[25]  Gabor T. Herman,et al.  Image Reconstruction From Projections , 1975, Real Time Imaging.

[26]  Weixue Lu,et al.  Multicriterion maximum entropy image reconstruction from projections , 1992, IEEE Trans. Medical Imaging.

[27]  C. A. Zala,et al.  Comparison of Algorithms for High Resolution Deconvolution of Array Beamforming Output , 1985 .

[28]  Azriel Rosenfeld,et al.  Digital Picture Processing , 1976 .

[29]  Stephen Grossberg,et al.  A massively parallel architecture for a self-organizing neural pattern recognition machine , 1988, Comput. Vis. Graph. Image Process..

[30]  Stuart Haber,et al.  A VLSI-efficient technique for generating multiple uncorrelated noise sources and its application to stochastic neural networks , 1991 .

[31]  Stephen Grossberg,et al.  Nonlinear neural networks: Principles, mechanisms, and architectures , 1988, Neural Networks.

[32]  C K Zoltani,et al.  Maximum bounded entropy: application to tomographic reconstruction. , 1985, Applied optics.

[33]  Bernard Widrow,et al.  30 years of adaptive neural networks: perceptron, Madaline, and backpropagation , 1990, Proc. IEEE.

[34]  G. Orlandi,et al.  Spectral estimation by repetitive L1-norm minimization , 1986, Proceedings of the IEEE.

[35]  James A. Cadzow,et al.  Functional analysis and the optimal control of linear discrete systems , 1973 .

[36]  Yoshiyasu Takefuji,et al.  Neural network parallel computing , 1992, The Kluwer International Series in Engineering and Computer Science.

[37]  Andrzej Cichocki,et al.  MOS Switched-Capacitor and Continuous-Time Integrated Circuits and Systems: Analysis and Design , 1989 .