Search

Here an image restoration on the basis of pixel simultaneous detection probabilities (PSDP) is proposed. These probabilities can be precisely determined by means of correlations measurement [NIMA 586 (2008) 314-326]. The proposed image restoration is based on the solution of matrix equation. Non-zero elements of Toeplitz block matrix with ones on the main diagonal, is determined using PSDP. The number of non zero descending diagonals depends on the detector construction and is not always smaller than 8. To solve the matrix equation, the Gaussian elimination algorithm is used. The proposed restoration algorithm is studied by means of the simulated images (with and without additive noise using PSDP for General Electric Senographe 2000D mammography device detector) and a small area (160 x 160 pixels) of real images acquired by the above mentioned device. The estimation errors of PSDP and the additive noise magnitude permits to restore images with the precision better than 3 % for the above mentioned detector. The additive noise in the real image is present after restoration and almost has the same magnitude. In the restored small area (16 x 16 mm) of real images, the pixel responses are not correlated. The spatial resolution improvement is also analyzed by the image of an absorber edge. Index Terms--Correlations, image restoration, pixel simultaneous detection probabilities, point spread function, Swank factor.

We propose a novel model for the measurement of a Mueller matrix for biological tissues. Compared with earlier measurement methods, our method can reduce measurement times and can significantly improve measurement efficiency. Our model needs only six intensity measurements to derive all 16 Mueller matrix components of a biological sample in arbitrary pixels. We used the pellicle cell of magnolia as our sample, and the experimental results are identical with those obtained with other methods. We demonstrate that we can obtain Mueller matrix components to recognize different biological tissues in the same visual field rapidly and reliably. OCIS codes: 120.5410, 180.0180.

We describe a novel approach developed for the National Polar-Orbiting Operational Environmental Satellite System/Visible Infrared Imaging Radiometer Suite (VIIRS) to retrieve pixel-level mixed-phase cloud optical thicknesses and effective particle sizes using 0.67, 1.6, 2.25, and 3.7 [micro]m bands reflectance/ radiance. This approach utilizes lookup tables of reflectances constructed from radiative transfer simulations and a numerical iterative search method. The capability of this new approach was demonstrated using Moderate Resolution Imaging Spectroradiometer (MODIS) data as proxy to VIIRS. Two proxy scenes, 14 October 2001 over North Platte, Nebraska, during the ninth Cloud Layer Experiment (CLEX-9) and 9 November 2006 over the Great Lakes and Eastern Canada during the Canadian CloudSat/Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations Validation Project (C3VP), were analyzed. The performance of the mixed-phase retrieval algorithm was assessed by comparison with the MODIS retrieval products, airborne in situ observations during CLEX-9 and CloudSat data during C3VP. OCIS codes: 010.1615, 280.1310.

We propose a spectral imaging method for piecewise 'macropixel' objects, which allows a regular digital camera to be converted into a digital snapshot spectral imager by equipping the camera with only a disperser and a demultiplexing algorithm. The method exploits a 'multiplexed spectrum' intensity pattern, i.e., the superposition of spectra from adjacent different image points, formed on the image sensor of the digital camera. The spatial image resolution is restricted to a macropixel level in order to acquire both spectral and spatial data (i.e., an entire spectral cubet in a single snapshot. Results of laboratory experiments with a special macropixel object image, composed of small, spatially uniform squares, provide to our knowledge a first verification of the proposed spectral imaging method. OCIS codes: 110.4234, 300.6190, 260.2030.

A region-based approach is presented for pixel clustering prior to the processes of attention in artificial vision systems. Results demonstrate the efficiency and robustness of the proposed technique.

The registration of multimodal images by using minimal spanning graphs as an alignment method is discussed. An efficient graph theoretic algorithm that estimates an alignment measure is presented.

CMOS Monolithic Active Pixel Sensors (MAPS) provide an attractive solution for high precision tracking of minimum ionizing particles. In these devices, a thin, moderately doped, undepleted silicon layer is used as the active detector volume with the readout electronics implemented on top of it. Recently, a new MAPS prototype was fabricated using the AMS 0.35 [micro]m OPTO process, featuring a thick epitaxial layer. A systematic study of tracking performance of that prototype using high-energy particle beam is presented in this work. Noise performance, signal amplitude from minimum ionizing particles, detection efficiency, spurious hit suppression and spatial resolution are shown as a function of the readout pitch and the charge collecting diode size. A test array with a novel readout circuitry was also fabricated and tested. Each pixel circuit consists of a front-end voltage amplifier, capacitively coupled to the charge collecting diode, followed by two analog memory cells. This architecture implements an on-pixel correlated double sampling method, allowing for optimization of integration independently of full frame readout time and strongly reduces the pixel-to-pixel output signal dispersion. First measurements using this structure are also presented. Index Terms--CMOS, monolithic, pixel, tracking.

We present a novel analog VLSI implementation of visual motion computation based on the lateral inhibition and positive feedback mechanisms that are inherent in the hysteretic winner-take-all circuit. By use of an input-dependent bias current and threshold mechanism, the circuit resets itself to prepare for another motion computation. This implementation was inspired by the Barlow-Levick model of direction selectivity in the rabbit retina. Each pixel uses 33 transistors and two small capacitors to detect the direction of motion and can be altered with the addition of six more transistors to measure the interpixel transit time. Simulation results and measurements from fabricated VLSI designs are presented to show the operation of the circuits. Index Terms--Analog VLSI, motion, pixel parallel, vision chips, VLSI design.

The exact relationship between the continuous and the discrete phase of two shifted images is established and it is shown that their discrete phase difference is a two-dimensional sawtooth signal. Subpixel registration can thus be performed directly in the Fourier domain by counting the number of cycles of the phase difference matrix along each frequency axis, and the solution is robustified by assuming a mixture model of inlying and outlying estimated shifts.

Contemporary event building approaches use data switches, either homemade or commercial off-the-shelf ones, to merge data from different channels and distribute them among processor nodes. However, in many trigger and DAQ systems, the merging and distributing functions can often be performed in pre-processing stages. By carefully integrating these functions into the upstream pre-processing stages, the events can be built without dedicated switches. In addition to the cost reducing, extra benefits are gain when the event is built early upstream. In this document, an example of the integrated upstream parasitic event building architecture that has been studied for the BTeV level 1 pixel trigger system is described. Several design considerations that experimentalists of other projects might be interested in are also discussed. Index Terms--DAQ, event building, trigger.

This paper presents a novel and efficient texture-based method for modeling the background and detecting moving objects from a video sequence. Each pixel is modeled as a group of adaptive local binary pattern histograms that are calculated over a circular region around the pixel. The approach provides us with many advantages compared to the state-of-the-art. Experimental results clearly justify our model. Index Terms--Motion, texture, background subtraction, local binary pattern.

In this paper, we introduce new types of variational segmentation cost functions and associated active contour methods that are based on pairwise similarities or dissimilarities of the pixels. As a solution to a minimization problem, we introduce a new curve evolution framework, the graph partitioning active contours (GPAC). Using global features, our curve evolution is able to produce results close to the ideal minimization of such cost functions. New and efficient implementation techniques are also introduced in this paper. Our experiments show that GPAC solution is effective on natural images and computationally efficient. Experiments on grayscale, color, and texture images show promising segmentation results. Index Terms--Curve evolution, active contours, image segmentation, pairwise similarity measures, graph partitioning.

A DePMOS structure provides detection and amplification jointly, and it is free of interconnection stray capacitance. To fully exploit the intrinsic low noise of the device an electrical model has been developed. The most relevant parameters have been measured in order to choose an adequate readout electronics. DePMOS can operate in continuous mode, i.e., without applying any clear pulse during the signal processing, and can be read out by a time-continuous shaper amplifier. An unequalled noise of 2.2 electrons rms at room temperature has been measured. In this mode DePMOS can be used, for example, as the readout device for silicon drift detectors. Anyway, DePMOS was developed to be the basic element of an active pixel sensor suitable to cope with the requirements of the XEUS Wide Field Imager. In a matrix arrangement, each pixel must be read out by a time-variant filter. A multichannel integrated shaping amplifier, based on multi-correlated double sampling, has been designed. Spectroscopic resolution obtained filtering the pixel matrix with this readout chip is in agreement with measurements in continuous mode and matches the predictions of the model presented. It has also been experimentally proved that the clear procedure does not introduce additional noise contribution, even in the very low noise range achieved. This qualifies DePMOS as a 'reset-noise-free' device. Index Terms--Active pixel sensor (APS), DePMOS, noise filtering, readout electronics, X-ray detector, X-ray Evolving Universe Spectroscopy (XEUS), X-ray spectroscopy.

We present a compact and low power integrated circuit designed to control the reset and perform pulse pile-up rejection in multi-channel spectroscopic-grade ASICs. The circuit has been implemented in a 0.35[micro]m CMOS technology using an area of 60 x 80 [micro][m.sup.2] and null static power consumption. These features make this circuit suitable to be embedded into the front-end readout cells for spectroscopy/imaging X- and gamma-ray pixel detectors. Index Terms--Application specific integrated circuit, front end electronics, pixel detector, semiconductor detector, X-ray detector.

A novel wide dynamic range (WDR) snapshot active pixel sensor for ultra-low power applications is presented. The proposed imager allows capturing of fast moving objects in the field of view and provides WDR by applying adaptive exposure time to each pixel, according to the local illumination intensity level. Driven by low-power dissipation requirements, the proposed pixel is operated by dual low voltage supplies (1.2 and 1.8 V) and utilizes an advanced low-power sensor design methodology. A test chip of a 32 * 32 array has been implemented in a standard 0.35-[micro]m CMOS technology. A single pixel occupies 18 * 32 [micro]m area and is expected to dissipate 18.5 nW at video rate. System architecture and operation are discussed and simulation results are presented. Index Terms--Active pixel sensor (APS), low-power sensor, snapshot CMOS imager, wide dynamic range (WDR) sensor.

Abstract--We introduce an effective approach for automatic offline authentication of handwritten samples where the forgeries are skillfully done, i.e., the true and forgery sample appearances are almost alike. Subtle details [...]

An adaptive partial disorder search algorithm entitled as the CPME-PDS (clustered pixel matching error-partial distortion search) is proposed. The algorithm makes use of the phenomenon that pixel matching errors in a macroblock with similar magnitude seem to appear together in a cluster in natural video sequences.

Problems such as preservation of discontinuities, model/data errors, outliers, confidence measures and performance evaluation in optical flow within a new framework are addressed. A blind adaptive technique based on generalized cross validation for estimating an independent regularization parameter for each pixel is derived.

A new class of related algorithms for deblocking block-transform compressed images and video sequences is proposed. The algorithms apply weighted sums on pixel quartets, which are symmetrically aligned with respect to block boundaries.

The development of an ultralow dark-signal and high-sensitivity pixel for an embedded active-pixel CMOS image sensor is described. A combined photogate/photodiode photon-sensing device with a novel operation scheme is developed to achieve in-pixel dark-current cancellation.

Often, in synthetic aperture radar (SAR) images of polar ice, one encounters shadow-like features across the images. Such features make it difficult to classify pixels into ice and water. Accordingly, it becomes a challenge to determine the true size and boundaries of ice floes in an SAR image of polar ice. In this paper, we develop a simple statistical procedure which classifies pixels of an image by eliminating the effects of shadow-like features. Methodology developed in this paper is illustrated using some noisy SAR images of ice floes in the Arctic sea. Index Terms--Binarization, E-M algorithm, Monte Carlo Markov chain (MCMC), shadow-like features.

A high dynamic range (HDR) CMOS image sensor architecture based on spiking pixel cells and its first implementation, a linear 128-pixel sensor, is presented. The pulse rate of the pixei cell is proportional to the local light intensity. The pixel cell features a pixel-level analog-to-digital (A/D) conversion by pulse counting. It allows to tradeoff conversion precision versus frame rate. Sample sensor chips have been manufactured using a 0.8-[micro]m double metal CMOS technology, and the test results confirming the feasibility of the chosen approach are reported. Index Terms--CMOS image sensors, high dynamic range, high value resolution, pixel-level analog-to-digital (A/D) conversion.

An ionizing radiation-tolerant CMOS active pixel sensor (APS) image sensor test chip was designed employing the physical design techniques of enclosed geometry and P-channel guard rings. The test chip was fabricated in a standard 0.35-[micro]m CMOS process that has a gate-oxide thickness of 7.0 nm. It was irradiated by a [gamma]-ray source up to a total ionizing radiation dose level of approximately 30 Mrd (Si) and was still functional. The most pronounced effect was the increase of dark current, which was linear with total dose level. The rate of dark current increase was about 1 to 2 pA/[cm.sup.2]/Krd (Si), depending on the design of the pixel. The results demonstrate that CMOS APS image sensors can be designed to be ionizing radiation tolerant to total dose levels up to 30 Mrd (Si). The fabrication process is standard CMOS, yielding a significant cost advantage over specialized radiation hard processes. Index Terms--Active pixel sensor, CMOS active pixel sensor, CMOS APS, dark current, image sensor, imager, ionizing radiation, ionizing radiation-induced dark current, ionizing radiation-tolerant CMOS APS, radiation hard, radiation tolerant.

A unified model is described for the overall modulation transfer function (MTF) estimation for complementary metal-oxide semiconductor (CMOS) image sensors. The model deals with the physical-diffusion effect, as well as the influence of the pixel active area geometrical shape.

The dark current of pixels in a charge-coupled device chip were studied in terms of the Meyer-Neldel rule.

A new study investigates the compaction of ordered dithered images using arithmetic coding.

Pixel voltage error was simulated for amorphous-silicon thin-film transistors.

Performance for a color CMOS photogate image sensor has been that found comparable to low-end charge-coupled device (CCD) sensors, but less effective than high-end ones. There is potential for a serious challenge to CCD by cancelling all fixed-pattern noise from read-out circuit mismatch using two levels of correlated-double sampling

We show that a pixelated lens with appropriate parameters exhibits an apodized point-spread function that originates in the finite size of the pixel's pupil. We evaluate numerically the degree of apodization and the enlargement associated with the point-spread function in terms of the parameters that characterize the pixelated lens.

A compact two-step modified-signed-digit arithmetic-logic array processor is proposed. When the reference digits are programmed, both addition and subtraction can be performed by the same binary logic operations regardless of the sign of the input digits. The optical implementation and experimental demonstration with an electron-trapping device are shown. Each digit is encoded by a single pixel, and no polarization is included. Any combinational logic can be easily performed without optoelectronic and electro-optic conversions of the intermediate results. The system is compact, general purpose, simple to align, and has a high signal-to-noise ratio.

A general purpose gamma ray camera has been designed and tested by Monte Carlo simulation. The design uses long thin fingers of LSO to obtain good spatial resolution at gamma ray energies from 100 keV to 100 MeV. The simulation was validated by comparing it with measurements on individual LSO fingers measuring 4x4x10mm and 4x4x75mm. Depth-of-interaction encoding, required for light-loss compensation, is obtained by a simple modification of the reflective coating of the fingers.

An enhanced nonnegativity and support constraints recursive inverse filtering (NAS-RIF) algorithm for blind image deconvolution is proposed. To overcome the problem of operation on images with different scales for the representation of pixel intensity levels, the original cost function is modified. The convergence of the conjugate gradient algorithm is improved using algorithm resetting. Results demonstrate the performance of the resulting enhanced NAS-RIF algorithm on various images.

Interferometric synthetic aperture radar (InSAR) ability to determine pixel elevation with a resolution derived from the interferometer phase resolution, is maintained as long as the interferometer phase change over the pixel does not exceed the phase resolution. This work investigates the surface coverage where this condition is met, determines the elevation resolution, and derives the ratio between elevation resolution and across-track resolution.

We demonstrate a method of performing the absolute three-flat test by using reflection symmetries of the surfaces and an algorithm for generating the rotation of arrays of pixel data. Most of the operations involve left/right and top/bottom flips of data arrays, operations that are very fast on most frame grabbers and are available on most commercial phase-measuring interferometers. We demonstrate the method with simulated data as well as with actual data from 150-mm-diameter surfaces that are flat to less than 25 nm peak to valley.

A study was conducted on fuzzy homogeneity vectors for handling the grayness and spatial uncertainties of pixels and to conduct multilevel thresholding. The spatial ambiguity among pixels exhibits inherent vagueness rather than randomness thus making conventional techniques inappropriate. The findings demonstrate that the proposed technique performs better than the histogram-based algorithms.

A study was conducted on a spatially variant finite mixture model for pixel labeling and image segmentation. An expectation-maximization (EM) algorithm has been derived for maximum likelihood estimation of the pixel labels and the parameters of the mixture densities. The proposed model can incorporate a spatial variation in the mixture weights assigned to different pixel locations. Moreover, the EM algorithm for the Bayesian estimation of pixel label parameters combines experimental knowledge from the observations and the researcher's insight in the form of prior densities on the label parameters.

An orthogonal subspace projection (OSP) method using linear mixture modeling was recently explored in hyperspectral image classification and has shown promise in signature detection, discrimination, and classification. In this paper, the OSP is revisited and extended by three unconstrained least squares subspace projection approaches, called signature space OSP, target signature space OSP, and oblique subspace projection, where the abundances of spectral signatures are not known a priori but need to be estimated, a situation to which the OSP cannot be directly applied. The proposed three subspace projection methods can be used not only to estimate signature abundance, but also to classify a target signature at subpixel scale so as to achieve subpixel detection. As a result, they can be viewed as a posteriori OSP as opposed to OSP, which can be thought of as a priori OSP. In order to evaluate these three approaches, their associated least squares estimation errors are cast as a signal detection problem in the framework of the Neyman-Pearson detection theory so that the effectiveness of their generated classifiers can be measured by characteristics (ROC) analysis. All results are demonstrated by computer simulations and Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) data. Index Terms - Classification, detection, hyperspectral image, oblique subspace projection classifier (OBC), orthogonal subspace projection (OSP), receiver operating characteristics (ROC), signature space orthogonal projection classifier (SSC), target signature space orthogonal projection classifier (TSC).

We present an architecture in which a multichannel correlator can perform simultaneous optical pattern recognition. Processing in parallel is made possible by use of the different diffraction orders produced by the pixelated structure of the liquid-crystal spatial light modulator employed to display the input scene. We codify additional quadratic phases in the filters to separate the correlation information corresponding to each channel. We demonstrate that the system can recognize different targets simultaneously. Good agreement between experimental and numerically simulated results is obtained. OCIS codes: 070.4550, 200.4960, 070.5010, 200.3050.

An analysis of the experimental results in the fabrication of complementary metal oxide semiconductor (CMOS) active pixel image sensors was presented. The CMOS technology, which was employed on the construction of the pixel sensors, was not modified to determine the impact of device scaling on the image sensing performance. The analysis revealed that the rise and fall transition times of output signal were long but the basic operations of the pixels were unaffected.

Because of image sampling, traditional measures of pixel dissimilarity can assign a large value to two corresponding pixels in a stereo pair, even in the absence of noise and other degrading effects. We propose a measure of dissimilarity that is probably insensitive to sampling because it uses the linearly interpolated intensity functions surrounding the pixels. Experiments on real images show that our measure alleviates the problem of sampling with little additional computational overhead. Index Terms - Dissimilarity, stereo matching, correspondence.

The spatially varying geometric response of the collimator-detector system in single photon emission computed tomography (SPECT) causes loss in resolution, shape distortions, reconstructed density nonuniformity, and quantitative inaccuracies. A projection space image reconstruction algorithm is used to correct these reconstruction artifacts. The projectors F use strip functions to calculate pixels more 'natural' for modeling the two-dimensional (2-D) geometric response of the SPECT collimator transaxially to the axis of rotation. These projectors are defined by summing the intersection of an array of multiple strips rotated at equal angles to approximate the ideal system geometric response of the collimator. Two projection models were evaluated for modeling the system geometric response function. For one projector each strip is of equal weight, for the other projector a Gaussian weighting is used. Parallel beam and fan beam projections of a physical three-dimensional (3-D) Hoffman brain phantom and a Jaszczak cold rod phantom were used to evaluate the geometric response correction. Reconstructions were obtained by using the singular value decomposition (SVD) method and the iterative conjugate gradient algorithm to solve for q in the imaging equation FGq = p, where p is the projection measurement. The projector F included the new models for the geometric response, whereas, the backprojector G did not always model the geometric response in order to increase the computational speed. The final reconstruction was obtained by sampling the backprojection Gq at a discrete array of points. Reconstructions produced by the two proposed projectors showed improved resolution when compared against a unit-strip 'natural' pixel model, the conventional image pixelized model with ray tracing to calculate the geometric response, and the filtered backprojection algorithm. When the reconstruction is displayed on fine grid points, the continuity and resolution of the image is preserved without the ring artifacts seen in the unit-strip 'natural' pixel model. With present computing power, the geometric response correction using the proposed projection space reconstruction approach is not yet feasible for routine clinical use. Index Terms - Collimator, geometric response, natural pixels, projection space image reconstruction, single value decomposition, SPECT.

Textures are frequently described using co-occurrence histograms of gray levels at two pixels in a given relative position. Analysis of several co-occurring pixel values may benefit texture description but is impeded by the exponential growth of histogram size. To make use of multidimensional histograms, we have developed methods for their reduction. The method described here uses linear compression, dimension optimization, and vector quantization. Experiments with natural textures showed that multidimensional histograms reduced with the new method provided higher classification accuracies than the channel histograms and the wavelet packet signatures. The new method was significantly faster than our previous one. Index Terms - Texture classification, multidimensional histograms, vector quantization, self-organizing map, feature selection.

The radiometric correction of synthetic aperture radar (SAR) imagery for topographically induced changes in the pixel scattering area is examined. The relative significance of ground slope in the range and azimuth directions is compared, and a widely used correction factor, based on the sin of the local radar incidence angle, is shown to be sufficiently accurate. The importance of performing this correction in slant range, rather than ground range, is demonstrated. Index Terms - AIRSAR, forest, scattering area normalization, synthetic aperture radar (SAR).

The independent pixel approximation (IPA), which treats radiative properties of each pixel independently by using standard plane-parallel calculations preserves scale-invariance found in the analyses of the horizontal variability of liquid water in marine stratocumulus clouds. Several studies, however, report a violation of scale-invariance in LANDSAT cloud radiance fields that are much smoother than cloud structure on small scales. This shows a limitation of IPA on small scales: it is unable to simulate the smooth small-scale behavior that is due to the horizontal photon transport. This paper introduces a 'nonlocal' independent pixel approximation (NIPA) that extends the IPA by incorporating empirically the smoothing effects of horizontal interpixel fluxes through a convolution product of the IPA and an approximate Green function for radiative transfer. We also address the inverse problem of cloud optical depth retrieval from satellite data, showing how NIPA can be used to overcome the limitations of current IPA-based methods at small scales.

A study was conducted of InP-InGaAs optoelectronic smart pixels for applications in optical interconnection and computing. The design, fabrication and performance of a high-sensitivity pixel and a high-bandwidth pixel were presented. The results indicate a minimum switching energy of 6 fJ, a maximum pixel bandwidth of 800 MHz and an optoelectronic gain of 3, the best overall performance characteristics of any optoelectronic smart pixel technology ever achieved.

Algebraic Reconstruction Technique (ART) is an age-old method used for solving the problem of three-dimensional (3-D) reconstruction from projections in electron microscopy and radiology. In medical applications, direct 3-D reconstruction is at the forefront of investigation. The simultaneous iterative reconstruction technique (SIRT) is an ART-type algorithm with the potential of generating in a few iterations tomographic images of a quality comparable to that of convolution-backprojection (CBP) methods. Pixel-based reconstruction (PBR) [10] is similar to SIRT reconstruction, and it has been shown that PBR algorithms give better quality pictures compared to those produced by SIRT algorithms. In this work, we propose a few modifications to the PBR algorithms. The modified algorithms are shown to give better quality pictures compared to PBR algorithms. The PBR algorithm and the modified PBR algorithms are highly compute intensive. Not many attempts have been made to reconstruct objects in the true 3-D sense because of the high computational overhead. In this study, we have developed parallel two-dimensional (2-D) and 3-D reconstruction algorithms based on modified PBR. We attempt to solve the two problems encountered by the PBR and modified PBR algorithms, i.e., the long computational time and the large memory requirements, by parallelizing the algorithm on a multiprocessor system. We investigate the possible task and data partitioning schemes by exploiting the potential parallelism in the PBR algorithm subject to minimizing the memory requirement. We have implemented an extended hypercube (EH) architecture for the high-speed execution of the 3-D reconstruction algorithm using the commercially available fast floating point digital signal processor (DSP) chips as the processing elements (PE's) and dual-port random access memories (DPR) as channels between the PE's. We discuss and compare the performances of the PBR algorithm on an IBM 6000 RISC workstation, on a Silicon Graphics Indigo 2 workstation, and on an EH system. The results show that an EH(3,1) using DSP chips as PE's executes the modified PBR algorithm about 100 times faster than an IBM 6000 RISC workstation. We have executed the algorithms on a 4-node IBM SP2 parallel computer. The results show that execution time of the algorithm on an EH(3,1) is better than that of a 4-node IBM SP2 system. The speed-up of an EH(3,1) system with eight PE's and one network controller (NC) is approximately 7.85.