Small convolution kernels for high-fidelity image restoration

An algorithm is developed for computing the mean-square-optimal values for small, image-restoration kernels. The algorithm is based on a comprehensive, end-to-end imaging system model that accounts for the important components of the imaging process: the statistics of the scene, the point-spread function of the image-gathering device, sampling effects, noise, and display reconstruction. Subject to constraints on the spatial support of the kernel, the algorithm generates the kernel values that restore the image with maximum fidelity, that is, the kernel minimizes the expected mean-square restoration error. The algorithm is consistent with the derivation of the spatially unconstrained Wiener filter, but leads to a small, spatially constrained kernel that, unlike the unconstrained filter, can be efficiently implemented by convolution. Simulation experiments demonstrate that for a wide range of imaging systems these small kernels can restore images with fidelity comparable to images restored with the unconstrained Wiener filter. >

[1]  Martin J. Lahart.  Local image restoration by a least-squares method , 1979 .

[2]  Carl L. Fales,et al.  Image-gathering system design for information and fidelity , 1988 .

[3]  R A Schowengerdt,et al.  Modulation-transfer-function analysis for sampled image systems. , 1984, Applied optics.

[4]  John A. Stuller.  An algebraic approach to image restoration filter design , 1972, Comput. Graph. Image Process..

[5]  Stephen E. Reichenbach,et al.  Digital image gathering and minimum mean-square error restoration , 1990, Other Conferences.

[6]  Charles R. Johnson,et al.  Matrix analysis , 1985, Statistical Inference for Engineers and Data Scientists.

[7]  F. O. Huck,et al.  Wiener restoration of sampled image data: end-to-end analysis , 1988 .

[8]  John A. Stuller,et al.  Transfer Function Compensation of Sampled Imagery , 1972, IEEE Transactions on Computers.

[9]  R A Schowengerdt,et al.  Image sampling, reconstruction, and the effect of sample-scene phasing. , 1982, Applied optics.

[10]  Clare McGillem,et al.  Optimum Constrained Image Restoration Filters , 1977, IEEE Transactions on Aerospace and Electronic Systems.

[11]  Friedrich O. Huck,et al.  Image gathering and restoration: information and visual quality , 1989 .

[12]  Stephen E. Reichenbach,et al.  Characterizing digital image acquisition devices , 1991 .

[13]  Stephen E. Reichenbach,et al.  Small-kernel image restoration , 1989 .

[14]  D. Graupe.  Image Restoration Filters Based on a 1-0 Weighting Over the Domain of Support of the PSF , 1979 .

[15]  James W. Modestino,et al.  Construction and properties of a useful two-dimensional random field , 1980, IEEE Trans. Inf. Theory.

[16]  C D McGillem,et al.  Constrained optimization of image restoration filters. , 1973, Applied optics.

[17]  G. F. Vermeij,et al.  The Approximation of Image Blur Restoration Filters by Finite Impulse Responses , 1980, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[18]  B E Saleh.  Trade off between resolution and noise in restoration by superposition of images. , 1974, Applied optics.

[19]  B. Roy Frieden.  Image restoration by discrete convolution of minimal length , 1974 .

[20]  F O Huck,et al.  Image gathering and processing: information and fidelity. , 1985, Journal of the Optical Society of America. A, Optics and image science.