Space-variant Blur Research Articles

MRF model-based identification of shift-variant point spread function for a class of imaging systems

Identification of pointwise space-variant blur is an important area of research in image processing and computer vision. In this paper, we propose an MAP-MRF based scheme for identifying a class of shift-variant imaging systems whose point spread function can be parameterized by a single blur parameter. The space-variant (SV) blur parameter is modeled as an MRF and the MAP estimate of the SV blur parameter is obtained using simulated annealing. The scheme is amenable to the incorporation of smoothness constraints on the spatial variations of the blur parameter and simultaneously allows for inclusion of line fields to preserve discontinuities in the shift-variant blur. The performance of the proposed scheme is tested on synthetic as well as real data and the estimates of the SV blur are found to be better than that of the window-based blur identification technique.

An MRF model-based approach to simultaneous recovery of depth and restoration from defocused images

In this paper, we propose a MAP-Markov random field (MRF) based scheme for recovering the depth and the focused image of a scene from two defocused images. The space-variant blur parameter and the focused image of the scene are both modeled as MRFs and their MAP estimates are obtained using simulated annealing. The scheme is amenable to the incorporation of smoothness constraints on the spatial variations of the blur parameter as well as the scene intensity. It also allows for inclusion of line fields to preserve discontinuities. The performance of the proposed scheme is tested on synthetic as well as real data and the estimates of the depth are found to be better than that of the existing window-based depth from defocus technique. The quality of the space-variant restored image of the scene is quite good even under severe space-varying blurring conditions.

Restoration of space-variant global blurs caused by severe camera movements and coordinate distortions

We show that a broad class of image recovery problems where an object undergoing an arbitrary two-dimensional, time- and space-variant, non-separable, nonlinear global coordinate distortion, is imaged for a certain duration, can be formulated as a system of linear inequalities. Since the system of inequalities arising in this context can be solved efficiently, our approach yields an effective method for solving this class of image recovery problems. A novel step size policy is introduced to accelerate the parallel surrogate constraint algorithm employed. The approach is illustrated by recovering an image severely blurred by the combined effects of translational and rotational motion and elliptic scaling. Resume. Nous demontrons qu'il est possible de reformuler un vaste groupe de problemes de recuperation d'image comme un systeme d'inegalites lineaires de grande dimension. Il s'agit en general, d'une image dans laquelle un objet subit pendant une certaine duree une distortion arbitraire, globale, nonseparable, nonlineaire de coordonnees en deux dimensions. Nous resolvons le systeme d'inegalites au moyen d'un nouvel algorithme de multiprojections simultanees. L'application de la methodologie est illustree sur l'exemple d'une image sujet a une distortion severe sous un mouvement de trajet complique. Mots cles: Recuperation d'image, restauration d'image, systemes d'inegalites lineaires, algorithmes de multiprojections

Parallel detection algorithm for page-oriented optical memories

We present a parallel algorithm for the reliable detection of two-dimensional binary data in page-oriented memories. The development of the proposed pseudodecision-feedback equalization (PDFE) method is motivated by the classical decision-feedback equalization receiver. The technique takes advantage of the known or the estimated optical system characteristics to mitigate space-variant blur and additive thermal noise. We extend the method to correct for fixed-pattern errors including magnification, rotation, and transverse shift. Advantages of the PDFE algorithm include its parallel design, low computational complexity, and local connectivity. A system-capacity metric is used to compare the performance of the PDFE receiver with other conventional approaches, including the simple threshold, the 1:2 modulation code, and the Wiener filter. Results show the PDFE to outperform all the above techniques over a variety of channels for both incoherent and coherent systems. Implementation issues are discussed, and a MOSIS (Metal-Oxide Semiconductor Implementation Service) 2-mum design is presented.

Blind restoration of images degraded by space-variant blurs using iterative algorithms for both blur identification and image restoration

This paper addresses the problem of digital blind restoration of images degraded by space-variant blurs and noise. The existing Expectation-Maximization (EM) algorithm reported in the literature is extended in this paper and combined with the region adaptive technique to handle the problem of identifying spatially variant blurs. The proposed algorithm is a two-step interative process. The expectation step of the EM algorithm is modified by the use of iterative image restoration. The entire image is divided into disjointed regions and the blur is identified in these regions using the proposed modified form of the EM algorithm. The iterative Constrained Least Squares (CLS) algorithm used in space-invariant image restoration is extended to restore the space variant blur images. Spatially adaptive algorithms for restoration are also applied. Experiments have been carried out to evaluate the performances of the proposed algorithms.

Scanning singular-value-decomposition method for restoration of images with space-variant blur

We present a method that can efficiently restore large images, blurred possibly nonuniformly and contaminated with noise, by use of a scanning singular-value-decomposition (SVD) method. Such an approach bypasses the prohibitive storage and speed limitations of the SVD method, thus, to our knowledge for the first time, making possible the restoration of reasonably sized images. We make use of the linear and local nature of the point spread function (PSF) to scan the image and restore it in the same raster without incurring blocking effects that are due to the overlap in neighboring reconstruction areas. The increase in speed compared with the conventional SVD approach can be many orders of magnitude, depending on the ratio of the point-spread blur to the image size. For example, if the linear extent of the PSF is one-eighth that of the image, a speed-up factor greater than 106 is achieved. A similar but less accurate solution to the problem of spatially variant blur by use of scanning Fourier transforms, which allows an even faster solution, is also described.

An image restoration system for a single-CCD color camcorder

A new image restoration system for improving the image of a single-CCD color camcorder is proposed. The single-CCD color camcorder reconstructs red, green, and blue color channels from a color filter array followed by a CCD. During this reconstruction process, color cross-talk among channels (interchannel distortion) and periodically space-varying blur (intrachannel distortion) occur. The proposed restoration system reduces interchannel distortion, and then, restores each color channel by removing the corresponding intrachannel distortion. Experimental results show that the proposed system provides the improved images in both objective and subjecture senses. The proposed system can be implemented by using a finite impulse response (FIR) filter structure, which is particularly suitable for real-time image processing. >

Phase-diversity correction of turbulence-induced space-variant blur

Space-variant blur is encountered when objects extend beyond the isoplanatic patch associated with the intervening atmospheric turbulence. The method of phase diversity, used to estimate jointly the object and the aberrations, is generalized to accommodate turbulence-induced space-variant blur. This generalization utilizes a parametric model for the blur function that is constructed with multiple phase screens. Simulation results are presented that demonstrate the recovery of near-diffraction-limited imagery from phase-diversity imagery that has been degraded with rather severe anisoplanatism.

Model-based segmentation and space-variant restoration of blurred images by decision-directed filtering

A new decision-directed filtering algorithm has been developed for model-based segmentation and space-variant restoration of blurred images. We assume that the space-variant blur can be represented by a collection of L distinct point spread functions, where L is a predetermined integer, such that at each pixel one of the L point spread functions will be more-or-less matched to the observed data. A multiple-model Kalman filtering procedure with on-line blur-model detection based on maximum a posteriori probability (MAP) decision was used to restore the image. The result of the decision process constitutes a model-based segmentation of the degraded image into regions of distinct blurs. There are several applications of the proposed algorithm. Among these, automatic blur detection, i.e., segmentation of a partially blurred image into regions of blur and no-blur is demonstrated as an example.

Some practical aspects of moving object deblurring in a perspective plane

An image restoration problem commonly encountered in many branches of science and engineering is that of removing the effects of uniform object motion. If the direction of object motion is not orthogonal to the line of sight of the imaging device, an image with both nonuniform motion blur and perspective may be produced so that the usual assumption of a space-invariant linear operation is no longer valid. In this paper we show that a proper initial geometrical transformation reduces the problem to one which can be handled by standard methods for removing uniform motion blur. As will be demonstrated, the process is strictly valid for planar (2-D) objects only. For objects with depth, residual space-variant blurs still exist. We highlight through examples several practical aspects of digitally deblurring such objects, including proper geometrical transformations, initial blur estimation, and residual space variance.

Image restoration of space-variant blurs by sectioned methods

Previous work on sectional methods in image processing is extended to the processing of degradations produced by space-variant point spread functions.

Space-variant system analysis of image motion*

A complete analysis of motion degradation in two-dimensional linear incoherent optical systems is presented from a system viewpoint. An equivalent space-variant point-spread function is derived from knowledge of the system motion, and the special characteristics of space-invariant and one-dimensional systems are considered. The analysis includes the effects of two different models for geometrical distortion in motion-degraded imaging and derives the distortions for aerial imaging. It is shown that a large class of space-variant operations may be decomposed into the cascade of invertible coordinate transformations and space-invariant systems; a block-diagram representation of this technique is presented. A number of examples of common types of motion blur are included, as well as computer simulations of space-variant blur and the decomposition process.