FOR EFFICIENT IMAGE PROCESSING. Hong Tang, Bingbing Zhou, Iain Macleod, Richard Brent and Wei Sun

Size: px
Start display at page:

Download "FOR EFFICIENT IMAGE PROCESSING. Hong Tang, Bingbing Zhou, Iain Macleod, Richard Brent and Wei Sun"

Transcription

1 A CLASS OF PARALLEL ITERATIVE -TYPE ALGORITHMS FOR EFFICIENT IMAGE PROCESSING Hong Tang, Bingbing Zhou, Iain Macleod, Richard Brent and Wei Sun Computer Sciences Laboratory Research School of Information Sciences and Engineering Australian National University Canberra, ACT 0200, Australia Abstract This paper proposes a class of parallel iterative median-type lters to restore signals degraded by Gaussian noise. An ecient implementation of these lters on distributed memory MIMD machines is outlined. Iterative median-type algorithms have been implemented with high eciency (> 90%) on the Fujitsu AP1000. We conclude that algorithms of this type are very suitable for parallel implementation on distributed memory machines. 1. Introduction Signal restoration, especially image restoration, requires large-volume and high-speed computations. Certain classes of image restoration operations such as spatial frequency ltering and convolution have relatively straightforward parallel implementations based on 2-D FFTs [1, 4, 5, 7, 11]. This paper discusses the more dicult case of non-linear ltering operations, with reference to application of a class of iterative median-type lters [2, 3, 4, 6, 8, 9, 10, 15] for enhancement of images degraded by additive Gaussian noise. The conventional median lter has the advantages that each point in the output image depends only on a limited neighbourhood of points in the input image (simplifying parallel implementation) and that it tends to preserve the shape of details in the original undegraded image. However, a conventional median lter usually needs to be applied iteratively (i.e. sequentially) to achieve the desired level of noise reduction. p-recursive median lters have better noise performance, but can lead to distortion of image detail. They also have a strong sequential aspect in that when computing a new output point they rely on values computed in the preceding p steps. One approach to parallel implementation of p-recursive median ltering divides the original image into sub-image blocks which are processed concurrently, but this is at the expense of introducing artefacts in the vicinity of the block boundaries. The following sections describe a hybrid parallel image lter, which combines p-recursive and conventional median lters in a way which gives superior image restoration (in terms of noise reduction, preservation of image detail and elimination of artefacts) and reduces overall computational and communication demands. 2. Median-Type Filters We discuss application of the standard median and the p-recursive median lters to restoration of noisy images. In a 2-D standard median lter with a square window of size (2k + 1) (2k + 1), the median replaces the central pixel: ^f(i; j) = MED f g(m; n) j i? k m i + k; j? k n j + kg (1) where ^f is the ltered image, g is the noisy image, and MED is the median operator. When the median lter is modied as follows, it becomes a \p-recursive median lter." The p-recursive median lter incorporates previous output values into the median decision process so that, at each step, it replaces the p points preceding the central point in the moving window with the medians from the previous p steps. The median operation can then be applied to obtain the new central value ^f(i; j). If the points in the window are numbered row by row in sequence, so that the central point is numbered l, where l = ((2k + 1) 2 + 1)=2, then the new method gives ^f l = MEDfg 1 ; :::; g l?p?1 ; ^f l?p ; :::; ^f l?1 ; g l ; :::; g (2k+1) 2g (2) where, for l? p r l? 1, ^f r is the median value of the window centred on the r th point. The parameter p can be changed to yield dierent lter types. The window moves from left to right and from top to bottom. Both standard median and p recursive median lters have good noise attenuation characteristics but at the expense of a possibly unacceptable amount of blurring. If p = 0, then Eq. (2) is the standard median lter. If p = 1, then there is only one output-feedback median point in the window. This has the eect of some noise reduction together with a very small amount of smoothing compared to the standard median lter. If p = l? 1, then Eq. (2) describes the recursive median lter [3, 8, 9] which has the greatest number of feedback median points in the window.

2 This provides much greater noise reduction than the rst case (p = 1), but usually results in an unacceptable degree of low pass ltering. We can generally choose an intermediate value of p (1 < p < l? 1) which gives an acceptable trade-o between noise reduction and loss of ne detail. When p = k(k + 2) = (k + 1) 2? 1; (3) for window size (2k + 1) (2k + 1), k=f1,2,3,...g, then the lter becomes the \improved" median lter [13, 14]. For a 3 3 window (k = 1), a good value of p is 3; for a 5 5 window (k = 2), p = 8; and so on. 3. Iterative Enhancement A simple model of image degradation is g(i; j) = f(i; j) + n(i; j); (4) where g(i; j); f(i; j) and n(i; j) are the degraded and original image and the noise respectively. The aim of image enhancement is to reduce the noise as much as possible or to nd a method which, given g, derives an image as close as possible to the original, subject to a suitable optimality criterion. A number of approaches can be found in the literature for solving the image enhancement problem [12]. The basic solution to Eq. (4) is as follows: ^f = D[g] = D[f + n]; (5) where D is an image enhancement lter. When D is of higher order, it becomes an iterative lter: f k+1 = D[f k ] (6) f 1 = D[f 0 + n] = D[f + n]; (7) where Eq. (7) is equal to Eq. (5). Iterative median lters are applied repeatedly to suppress noise signals while preserving image edges, according to a suitable criterion. Although they tend to be computationally expensive, such lters are very popular in image processing because of their good performance. They would nd widespread use in interactive applications if they could be made to operate in real-time: the most promising approach here is through parallel implementation. 4. Parallel Algorithms This section describes parallel implementation of iterative median lters on distributed memory MIMD machines (these having better scalability than shared memory machines). The implementations given are also suitable for SIMD machines with appropriate conditional operators. Assume that P processing elements (PEs) are available in the parallel system. To process an N N image we rst divide the image into P x P y ; (8) blocks, where P x ; P y 1 and P x P y = P, letting each block hold N P x P y adjacent pixels. Each sub-image block is then allocated to the local memory of a separate PE, so that the image is evenly distributed across the entire system. As described above, the new value of the central pixel depends on all the pixels in the processing window of a median lter. If the window size is M M, K = (M?1)=2 rows or columns of pixels at each boundary of a block will aect and also be aected by the pixels close to them in the original image but forming part of the boundary region of other blocks. Since the image is distributed across the system, we need to access the local memories of other PEs by using explicit data communication when the new values of boundary region pixels in each PE are being calculated. 4.1 Parallel implementation of median lter With median ltering the output image values from the current iteration depend only on the values from the previous iteration. We may thus combine all the required pixels at each boundary of a block and transfer them as a single message between the corresponding PEs at the beginning of each iteration. The pixels received in each PE pad the corresponding boundaries of each block. We then work only with such locally augmented blocks and there is no further data communication required during each iteration. We give an example here of a two step procedure for communicating data between PEs when the parallel computing system is organized as a 2-D array. In the rst step, the K boundary rows of pixels from the top and bottom boundaries of a block are transferred vertically between neighbouring PEs; the received rows pad the top and bottom of the block. In the second step, the K boundary columns from the (padded) left and right boundaries are transferred horizontally between the neighbouring PEs. Note that transfer of these padded columns also transfers the required elements from diagonally neighbouring PEs. As discussed above, the original N N image is divided into P x P y blocks and evenly distributed across a system of P = P x P y PEs. Each block is of size N=P x N=P y and its shape varies with the values of P x and P y. It is easy to show that the total number of pixels to be transferred

3 from other PEs to a given PE is about N 2K + N : (9) P y P x In the system-wide sense, this number can be minimised by choosing similar values for P x and P y so that the block sub-images are approximately square. Minimising communication overheads is important with distributed memory architectures because transfer of data elements between PEs is typically much slower than within PEs. 4.2 Parallel hybrid median lter Although they reduce noise more eciently than conventional median lters, p-recursive median lters cannot be directly implemented in parallel because the image values to be determined rely on the feedback of median points computed from the previous windows in the same iteration. This section describes a hybrid median lter which combines p-recursive and conventional median lters in a way which gives superior image restoration, allows parallel implementation and reduces overall computational cost. Suppose that the image is divided into blocks in a similar manner to that described above for the conventional median lter. Our parallel algorithm has two stages. In the rst stage, K 0 ( K) boundary columns or rows of each block are transferred to adjacent blocks and are used to pad the corresponding boundaries. The p-recursive median ltering algorithm is then applied to each augmented block, using only pixels local to each PE. All blocks can be processed in parallel with this block expanding method. As a result of regarding each augmented block as an independent image, artefacts are introduced in the vicinity of the block boundaries. Two techniques which can be combined to reduce such boundary eects to acceptable levels are now considered. When the columns or rows are exchanged between adjacent blocks, the amount of information obtained from the neighbours depends on the extent of padding or augmentation. The boundary artefacts are reduced in magnitude as the width of overlap K 0 increases. However, if we make the overlap large the total commmunication and computational costs increase substantially. Fortunately, our experimental results (discussed in the next section) show that in practice the magnitude of boundary artefacts falls quite quickly with increasing K 0 (for K 0 K). The second technique for reducing boundary eects has a more subtle basis. Observe that there will be no boundary eects if each pixel has a signicant correlation only with those pixels which reside in the same block. For most classes of imagery, the degree of correlation between pairs of pixels depends on their physical separation. Thus for a rectangular block with a given number of pixels, the mutually relevant information regarding pixels within the block is maximised if the block is square. Thus when P x and P y are similar, the total number of pixels to be transferred between blocks and the magnitude of boundary artefacts are both reduced. Even with use of the above techniques, noticeable nonconstant gray level dierences may still exist in the vicinity of block boundaries. A conventional median lter can then be applied to smooth the entire image and to eliminate remaining visible artefacts, forming the second stage of our parallel hybrid algorithm. 5. Experimental Results We report the results of implementation experiments on a practical MIMD machine, comparing the restoration performance of various lters on representative degraded images. Parallel Machine: Parallel algorithms for various median-type lters have been implemented on the ANU's Fujitsu AP1000, a distributed memory MIMD machine with 128 (1024 max.) SPARC processors. Each processor has 16 MBytes of memory and a Weitek oatingpoint unit. The topology of the AP1000 is a 2-D torus, with hardware support for wormhole routing. The routing hardware (T-net) provides a theoretical bandwidth of 25 MByte/sec between any two cells. In practice, about 6 MByte/sec is attainable by user programs. The routing hardware supports row and column broadcasts, with broadcast times being comparable to those of PE to PE transfers. Parallel Parameters: The two most important criteria for evaluating the success of the parallel implementations of the algorithms used in our experiments are speedup and eciency, dened as follows: Speedup: This is dened as the ratio between T 1, the time to execute the fastest serial algorithm for a given problem on one processor and T P, the time to execute the parallel algorithm for the same problem on P processors, that is, Speedup = T 1 =T P : The Speedup measures the overall improvement through parallelisation in solving a given problem. Eciency: This is dened as the Speedup divided by P, the total number of processors used in executing the parallel algorithm, that is, Eciency = Speedup=P: If the eciency is close to unity, it is clear that the processors are doing useful work most of the time and that communication and synchronisation overheads associated with the parallel implementation are relatively small. This means that the hardware is being used eectively. Image Data: Our experiments used three dierent image sources of varying sizes and characteristics. These

4 g(i,j) STANDARD ^ f(i,j) TABLE I The performance parameters of median-type filters in minimum CPU time and optimal image quality. g(i,j) Fig. 1. The median lter. RECURSIVE STANDARD ^ f(i,j) Fig. 2. Combined ltering by recursive and median lters in one iteration step. were \Girl" of size (smoothly varying photographic image), \Lenna" of size (many ne details and sharp edges) and \Christine" of size (large but relatively smooth). Gaussian noise was added to each image to give a signal-to-noise ratio (SNR) of about 10dB. Images of dierent sizes were used to study the way in which the speedup and eciency varied with problem size. Experimental Design: We have tested three types of iterative median lters with their parallel implementation. The performance of each of the lters and its combined forms with two of the other median lters have also been realised and given in Figures 1, 2 and 3. Their optimal combination regarding CPU time and image quality is given in Table I. From the table, it is clearly seen that parallel hybrid median lters greatly reduce the CPU times while still achieving the same improvement in signal-tonoise ratios. 5.1 Speedup and eciency of parallel lter implementations The overall results here were excellent, with speedups ranging from better than 15.5 with 16 processors up to 117 or so with 128 processors. The eciencies ranged from a low of 92% with 128 processors to nearly 100% with 16 processors. The measured eciencies were slightly better for the larger images. Parallel execution times ranged g(i,j) IMPROVED STANDARD ^ f(i,j) Fig. 3. Combined ltering by one iteration of improved median plus two iterations of median lter. Filter Iterations Total CPU S/N Gain sequential median sequential improved sequential recursive parallel improved+median parallel recursive+median from tenths of a second through to about ten seconds, compared with sequential times of from one to twenty minutes or so. The good results are in part a consequence of our data communication strategy which greatly reduces the total communication cost, so all the processors involved in the computation are active most of the time. 5.2 Image enhancement performance The experimental results conrmed previous observations [13, 14] that the conventional median lter requires more iterations to achieve a given gain in S/N ratio than either the recursive or improved median lters. Parallel implementations of recursive and improved median lters (based on the block expanded method) gave useful gains in overall S/N ratio with one or two iterations. Important requirements of image enhancement processes are that during noise reduction operations they should as far as possible preserve the sharpness and shape of detail in the original image and minimise visible artefacts such as \ringing." Repeated application of a conventional median lter to achieve the desired noise reduction tended to result in excessive blurring. Recursive median ltering gave good noise reduction without excessive blurring, but tended to distort high contrast image detail. A further problem was that in the parallel implementation of improved and recursive median lters, artefacts associated with block boundaries were distinctly visible with K = 1 (and less visible as K was increased). Hybrid lters based on one or more iterations of either an improved or recursive median lter, followed by two iterations of a conventional median lter, achieved useful S/N gains and were eective in reducing the visibility of block boundary artefacts. The measured overall gains in S/N ratio do not clearly indicate the reductions in artefact amplitude with such second stage ltering. To get a better measure of the improvements here, the gain in S/N ratio in narrow strips of the image centred on the block boundaries was calculated. This localised measurement was more sensitive to artefact amplitude and gave an objective assessment of the eect of increasing K and using

5 Distortion 10 (db) (a) Fig. 4. Signal-to-noise ratio versus K K second stage ltering. Figure 4 shows how the artefacts reduce with increasing K after parallel hybrid median ltering. Figure 5 gives an example of the ltered results using our lters on degraded 10dB Christine. 6. Conclusions The overall approach adopted here is to take an image processing procedure which has sequential aspects and to apply it in parallel to a number of sub-images, at the same time recognising that this may introduce artefacts in the vicinity of the sub-image boundaries. The magnitude of the artefacts can be reduced by analysing their causes and making suitable modications to the processing techniques. A nal ltering stage can then reduce any residual artefacts to insignicant levels. The parallel iterative median-type noise ltering algorithms presented in this paper are quite suitable for implementation on distributed memory MIMD and SIMD machines. The block expanded method allows all the required data to be transferred between adjacent PEs as single messages. This greatly reduces the overall communication overheads resulting from message latency. For a given number of processors we also show how to partition the original image so that the total amount of data to be transferred between PEs can be minimised without affecting the nal result, further reducing communication costs. High eciencies were achieved with the ANU's AP1000, indicating that the algorithms implemented are well matched to a mesh-connected processor array architecture. References [1] R. P. Brent, \Parallel algorithms for digital signal processing", Numerical Linear Algebra, Digital Signal Processing and Parallel Algorithms (edited by G. H. Golub and P. Van Dooren), Springer-Verlag, 1991, pp.93{110. [2] H. U. Dohler: \Generation of root signals of two dimensional median lters," Signal Processing, Vol.18, 1989, pp.269{276. [3] N. C. Gallagher and G. L. Wise: \A theoretical analysis of the properties of median lters," IEEE Trans. Acoust. Speech & Signal Process., Vol-29, 1981, pp.1136{1141. [4] K. S. Fu, VLSI for Pattern Recognition and Image Processing, Springer-Verlag, New York, (b) (c) (d) Fig. 5. The ltered results on degraded 10dB 1024X1024 Christine. (a) the degraded image; (b) after parallel median ltering; (c) after parallel hybrid improved median ltering (k = 1); (d) after parallel hybrid recursive median ltering (k = 1).

6 [5] L. Kronsjo and D. Shumsheruddin, Advances in Parallel Algorithms, Blackwell Scientic Publications, London, [6] T. S. Huang (Ed.): Two-Dimensional Digital Signal Processing II Spring-Verlag, Berlin Heidelberg, 1981 [7] A. Nakamura, M. Nivat, A. Saoudi, P. S. P. Wang and K. Inoue, Lecture Notes in Computer Science { Parallel Image Analysis, Springer-Verlag, 1992 [8] T. A. Nodes and N. C. Gallagher: \Median lters: Some modications and their properties," IEEE Trans. Acoust. Speech & Signal Process., Vol-30, 1982, pp.739{746. [9] I. Pitas and A. N. Venetsanopoulos: \Nonlinear mean lters in image processing", IEEE Trans. Acoust. Speech & Signal Process., Vol-34, 1986, pp.537{584 [10] A. Rosenfeld and A. C. Kak, Digital Picture Processing, Vols. 1 and 2, Academic Press [11] D. Rover, V. Tsai, Y. S. Chow and J. Gustafson: \Signalprocessing algorithms on parallel architectures: A performance update," Journal of Parallel and Distributed Computing, Vol- 13, 1991, pp.237{245. [12] M.Sarrafzadeh, A. K. Katsaggelos and S. P. R. Kumar, \Parallel architectures for iterative image restoration," in Parallel Algorithms and Architectures for DSP Applications, Academic Publishers, [13] Tang H., Zhou B.B., Macleod I.D., Brent R.P. and Sun W., \Comparisions of Parallel Iterative Noise Filters for Real-time Image Processing", to appear in the Proceedings of International Conference on Systems, Control, Information, Wuhan, China, October [14] Tang H., Zhou B.B., Macleod I.D., Brent R.P. and Sun W., \Parallel implementation of an adaptive dierential iterative RMS lter for ecient signal restoration", to appear in the Proceedings of IEEE Region 10 International Conference on Parallel Computation and Applications, Singapore, August [15] Tang H., \A human-machine interactive system for ecient image restoration", Proceedings of International Conference on Acoustics, Speech and Signal Processing 1994, Adelaide, Australia, April, Vol.5, 1994, pp [16] Tang H., \A combining approach for the enhancement of unknown blurred and noisy image", Proceedings of 1994 International Conference on Speech, Image and Neural Networks, Hong Kong, April, Vol.2, 1994, pp

Ecient Implementation of Sorting Algorithms on Asynchronous Distributed-Memory Machines

Ecient Implementation of Sorting Algorithms on Asynchronous Distributed-Memory Machines Ecient Implementation of Sorting Algorithms on Asynchronous Distributed-Memory Machines Zhou B. B., Brent R. P. and Tridgell A. y Computer Sciences Laboratory The Australian National University Canberra,

More information

Ecient Implementation of Sorting Algorithms on Asynchronous Distributed-Memory Machines

Ecient Implementation of Sorting Algorithms on Asynchronous Distributed-Memory Machines Ecient Implementation of Sorting Algorithms on Asynchronous Distributed-Memory Machines B. B. Zhou, R. P. Brent and A. Tridgell Computer Sciences Laboratory The Australian National University Canberra,

More information

Parallel Implementation of QRD Algorithms on the Fujitsu AP1000

Parallel Implementation of QRD Algorithms on the Fujitsu AP1000 Parallel Implementation of QRD Algorithms on the Fujitsu AP1000 Zhou, B. B. and Brent, R. P. Computer Sciences Laboratory Australian National University Canberra, ACT 0200 Abstract This paper addresses

More information

WEINER FILTER AND SUB-BLOCK DECOMPOSITION BASED IMAGE RESTORATION FOR MEDICAL APPLICATIONS

WEINER FILTER AND SUB-BLOCK DECOMPOSITION BASED IMAGE RESTORATION FOR MEDICAL APPLICATIONS WEINER FILTER AND SUB-BLOCK DECOMPOSITION BASED IMAGE RESTORATION FOR MEDICAL APPLICATIONS ARIFA SULTANA 1 & KANDARPA KUMAR SARMA 2 1,2 Department of Electronics and Communication Engineering, Gauhati

More information

Document Image Restoration Using Binary Morphological Filters. Jisheng Liang, Robert M. Haralick. Seattle, Washington Ihsin T.

Document Image Restoration Using Binary Morphological Filters. Jisheng Liang, Robert M. Haralick. Seattle, Washington Ihsin T. Document Image Restoration Using Binary Morphological Filters Jisheng Liang, Robert M. Haralick University of Washington, Department of Electrical Engineering Seattle, Washington 98195 Ihsin T. Phillips

More information

AN ALGORITHM FOR BLIND RESTORATION OF BLURRED AND NOISY IMAGES

AN ALGORITHM FOR BLIND RESTORATION OF BLURRED AND NOISY IMAGES AN ALGORITHM FOR BLIND RESTORATION OF BLURRED AND NOISY IMAGES Nader Moayeri and Konstantinos Konstantinides Hewlett-Packard Laboratories 1501 Page Mill Road Palo Alto, CA 94304-1120 moayeri,konstant@hpl.hp.com

More information

Network. Department of Statistics. University of California, Berkeley. January, Abstract

Network. Department of Statistics. University of California, Berkeley. January, Abstract Parallelizing CART Using a Workstation Network Phil Spector Leo Breiman Department of Statistics University of California, Berkeley January, 1995 Abstract The CART (Classication and Regression Trees) program,

More information

LINPACK Benchmark. on the Fujitsu AP The LINPACK Benchmark. Assumptions. A popular benchmark for floating-point performance. Richard P.

LINPACK Benchmark. on the Fujitsu AP The LINPACK Benchmark. Assumptions. A popular benchmark for floating-point performance. Richard P. 1 2 The LINPACK Benchmark on the Fujitsu AP 1000 Richard P. Brent Computer Sciences Laboratory The LINPACK Benchmark A popular benchmark for floating-point performance. Involves the solution of a nonsingular

More information

An Embedded Wavelet Video Coder. Using Three-Dimensional Set. Partitioning in Hierarchical Trees. Beong-Jo Kim and William A.

An Embedded Wavelet Video Coder. Using Three-Dimensional Set. Partitioning in Hierarchical Trees. Beong-Jo Kim and William A. An Embedded Wavelet Video Coder Using Three-Dimensional Set Partitioning in Hierarchical Trees (SPIHT) Beong-Jo Kim and William A. Pearlman Department of Electrical, Computer, and Systems Engineering Rensselaer

More information

Job Re-Packing for Enhancing the Performance of Gang Scheduling

Job Re-Packing for Enhancing the Performance of Gang Scheduling Job Re-Packing for Enhancing the Performance of Gang Scheduling B. B. Zhou 1, R. P. Brent 2, C. W. Johnson 3, and D. Walsh 3 1 Computer Sciences Laboratory, Australian National University, Canberra, ACT

More information

Extra-High Speed Matrix Multiplication on the Cray-2. David H. Bailey. September 2, 1987

Extra-High Speed Matrix Multiplication on the Cray-2. David H. Bailey. September 2, 1987 Extra-High Speed Matrix Multiplication on the Cray-2 David H. Bailey September 2, 1987 Ref: SIAM J. on Scientic and Statistical Computing, vol. 9, no. 3, (May 1988), pg. 603{607 Abstract The Cray-2 is

More information

An Embedded Wavelet Video. Set Partitioning in Hierarchical. Beong-Jo Kim and William A. Pearlman

An Embedded Wavelet Video. Set Partitioning in Hierarchical. Beong-Jo Kim and William A. Pearlman An Embedded Wavelet Video Coder Using Three-Dimensional Set Partitioning in Hierarchical Trees (SPIHT) 1 Beong-Jo Kim and William A. Pearlman Department of Electrical, Computer, and Systems Engineering

More information

ARTIFICIAL INTELLIGENCE LABORATORY. A.I. Memo No November, K.P. Horn

ARTIFICIAL INTELLIGENCE LABORATORY. A.I. Memo No November, K.P. Horn MASSACHUSETTS INSTITUTE OF TECHNOLOGY ARTIFICIAL INTELLIGENCE LABORATORY A.I. Memo No. 1584 November, 1996 Edge and Mean Based Image Compression Ujjaval Y. Desai, Marcelo M. Mizuki, Ichiro Masaki, and

More information

A New Orthogonal Multiprocessor and its Application to Image. Processing. L. A. Sousa M. S. Piedade DEEC IST/INESC. R.

A New Orthogonal Multiprocessor and its Application to Image. Processing. L. A. Sousa M. S. Piedade DEEC IST/INESC. R. A New Orthogonal ultiprocessor and its Application to Image Processing L. A. Sousa. S. Piedade email:las@molly.inesc.pt email: msp@inesc.pt DEEC IST/INESC R. Alves Redol,9 1000 Lisboa, PORTUGAL Abstract

More information

1 INTRODUCTION The LMS adaptive algorithm is the most popular algorithm for adaptive ltering because of its simplicity and robustness. However, its ma

1 INTRODUCTION The LMS adaptive algorithm is the most popular algorithm for adaptive ltering because of its simplicity and robustness. However, its ma MULTIPLE SUBSPACE ULV ALGORITHM AND LMS TRACKING S. HOSUR, A. H. TEWFIK, D. BOLEY University of Minnesota 200 Union St. S.E. Minneapolis, MN 55455 U.S.A fhosur@ee,tewk@ee,boley@csg.umn.edu ABSTRACT. The

More information

Controlling the spread of dynamic self-organising maps

Controlling the spread of dynamic self-organising maps Neural Comput & Applic (2004) 13: 168 174 DOI 10.1007/s00521-004-0419-y ORIGINAL ARTICLE L. D. Alahakoon Controlling the spread of dynamic self-organising maps Received: 7 April 2004 / Accepted: 20 April

More information

Fingerprint Image Enhancement Algorithm and Performance Evaluation

Fingerprint Image Enhancement Algorithm and Performance Evaluation Fingerprint Image Enhancement Algorithm and Performance Evaluation Naja M I, Rajesh R M Tech Student, College of Engineering, Perumon, Perinad, Kerala, India Project Manager, NEST GROUP, Techno Park, TVM,

More information

A Fourier Extension Based Algorithm for Impulse Noise Removal

A Fourier Extension Based Algorithm for Impulse Noise Removal A Fourier Extension Based Algorithm for Impulse Noise Removal H. Sahoolizadeh, R. Rajabioun *, M. Zeinali Abstract In this paper a novel Fourier extension based algorithm is introduced which is able to

More information

A Switching Weighted Adaptive Median Filter for Impulse Noise Removal

A Switching Weighted Adaptive Median Filter for Impulse Noise Removal A Switching Weighted Adaptive Median Filter for Impulse Noise Removal S.Kalavathy Reseach Scholar, Dr.M.G.R Educational and Research Institute University, Maduravoyal, India & Department of Mathematics

More information

Outline. Parallel Algorithms for Linear Algebra. Number of Processors and Problem Size. Speedup and Efficiency

Outline. Parallel Algorithms for Linear Algebra. Number of Processors and Problem Size. Speedup and Efficiency 1 2 Parallel Algorithms for Linear Algebra Richard P. Brent Computer Sciences Laboratory Australian National University Outline Basic concepts Parallel architectures Practical design issues Programming

More information

BIG DATA-DRIVEN FAST REDUCING THE VISUAL BLOCK ARTIFACTS OF DCT COMPRESSED IMAGES FOR URBAN SURVEILLANCE SYSTEMS

BIG DATA-DRIVEN FAST REDUCING THE VISUAL BLOCK ARTIFACTS OF DCT COMPRESSED IMAGES FOR URBAN SURVEILLANCE SYSTEMS BIG DATA-DRIVEN FAST REDUCING THE VISUAL BLOCK ARTIFACTS OF DCT COMPRESSED IMAGES FOR URBAN SURVEILLANCE SYSTEMS Ling Hu and Qiang Ni School of Computing and Communications, Lancaster University, LA1 4WA,

More information

Object classes. recall (%)

Object classes. recall (%) Using Genetic Algorithms to Improve the Accuracy of Object Detection Victor Ciesielski and Mengjie Zhang Department of Computer Science, Royal Melbourne Institute of Technology GPO Box 2476V, Melbourne

More information

Parallel Pipeline STAP System

Parallel Pipeline STAP System I/O Implementation and Evaluation of Parallel Pipelined STAP on High Performance Computers Wei-keng Liao, Alok Choudhary, Donald Weiner, and Pramod Varshney EECS Department, Syracuse University, Syracuse,

More information

On Parallel Implementation of the One-sided Jacobi Algorithm for Singular Value Decompositions

On Parallel Implementation of the One-sided Jacobi Algorithm for Singular Value Decompositions On Parallel Implementation of the One-sided Jacobi Algorithm for Singular Value Decompositions B. B. Zhou and R. P. Brent Computer Sciences Laboratory The Australian National University Canberra, ACT 000,

More information

Blur Space Iterative De-blurring

Blur Space Iterative De-blurring Blur Space Iterative De-blurring RADU CIPRIAN BILCU 1, MEJDI TRIMECHE 2, SAKARI ALENIUS 3, MARKKU VEHVILAINEN 4 1,2,3,4 Multimedia Technologies Laboratory, Nokia Research Center Visiokatu 1, FIN-33720,

More information

Hybrid filters for medical image reconstruction

Hybrid filters for medical image reconstruction Vol. 6(9), pp. 177-182, October, 2013 DOI: 10.5897/AJMCSR11.124 ISSN 2006-9731 2013 Academic Journals http://www.academicjournals.org/ajmcsr African Journal of Mathematics and Computer Science Research

More information

The 2D wavelet transform on. a SIMD torus of scanline processors. R. Lang A. Spray H. Schroder. Application Specic Computer Design (ASCOD)

The 2D wavelet transform on. a SIMD torus of scanline processors. R. Lang A. Spray H. Schroder. Application Specic Computer Design (ASCOD) The D wavelet transform on a SIMD torus of scanline processors R. Lang A. Spray H. Schroder Application Specic Computer Design (ASCOD) Dept. of Electrical & Computer Engineering University of Newcastle

More information

CHAPTER 4 AN INTEGRATED APPROACH OF PERFORMANCE PREDICTION ON NETWORKS OF WORKSTATIONS. Xiaodong Zhang and Yongsheng Song

CHAPTER 4 AN INTEGRATED APPROACH OF PERFORMANCE PREDICTION ON NETWORKS OF WORKSTATIONS. Xiaodong Zhang and Yongsheng Song CHAPTER 4 AN INTEGRATED APPROACH OF PERFORMANCE PREDICTION ON NETWORKS OF WORKSTATIONS Xiaodong Zhang and Yongsheng Song 1. INTRODUCTION Networks of Workstations (NOW) have become important distributed

More information

Noise filtering for television receivers with reduced memory

Noise filtering for television receivers with reduced memory Noise filtering for television receivers with reduced memory R. J. Schutten, G. de Haan and A. H. M. van Roermund. Philips Research Laboratories, Television Systems Group, Prof. Holstlaan 4, 5656 AA Eindhoven,

More information

sizes. Section 5 briey introduces some of the possible applications of the algorithm. Finally, we draw some conclusions in Section 6. 2 MasPar Archite

sizes. Section 5 briey introduces some of the possible applications of the algorithm. Finally, we draw some conclusions in Section 6. 2 MasPar Archite Parallelization of 3-D Range Image Segmentation on a SIMD Multiprocessor Vipin Chaudhary and Sumit Roy Bikash Sabata Parallel and Distributed Computing Laboratory SRI International Wayne State University

More information

Egemen Tanin, Tahsin M. Kurc, Cevdet Aykanat, Bulent Ozguc. Abstract. Direct Volume Rendering (DVR) is a powerful technique for

Egemen Tanin, Tahsin M. Kurc, Cevdet Aykanat, Bulent Ozguc. Abstract. Direct Volume Rendering (DVR) is a powerful technique for Comparison of Two Image-Space Subdivision Algorithms for Direct Volume Rendering on Distributed-Memory Multicomputers Egemen Tanin, Tahsin M. Kurc, Cevdet Aykanat, Bulent Ozguc Dept. of Computer Eng. and

More information

VIDEO DENOISING BASED ON ADAPTIVE TEMPORAL AVERAGING

VIDEO DENOISING BASED ON ADAPTIVE TEMPORAL AVERAGING Engineering Review Vol. 32, Issue 2, 64-69, 2012. 64 VIDEO DENOISING BASED ON ADAPTIVE TEMPORAL AVERAGING David BARTOVČAK Miroslav VRANKIĆ Abstract: This paper proposes a video denoising algorithm based

More information

PARALLEL COMPUTATION OF THE SINGULAR VALUE DECOMPOSITION ON TREE ARCHITECTURES

PARALLEL COMPUTATION OF THE SINGULAR VALUE DECOMPOSITION ON TREE ARCHITECTURES PARALLEL COMPUTATION OF THE SINGULAR VALUE DECOMPOSITION ON TREE ARCHITECTURES Zhou B. B. and Brent R. P. Computer Sciences Laboratory Australian National University Canberra, ACT 000 Abstract We describe

More information

Comparative Analysis in Medical Imaging

Comparative Analysis in Medical Imaging 1 International Journal of Computer Applications (975 8887) Comparative Analysis in Medical Imaging Ashish Verma DCS, Punjabi University 1, Patiala, India Bharti Sharma DCS, Punjabi University 1, Patiala,

More information

REDUCTION OF CODING ARTIFACTS IN LOW-BIT-RATE VIDEO CODING. Robert L. Stevenson. usually degrade edge information in the original image.

REDUCTION OF CODING ARTIFACTS IN LOW-BIT-RATE VIDEO CODING. Robert L. Stevenson. usually degrade edge information in the original image. REDUCTION OF CODING ARTIFACTS IN LOW-BIT-RATE VIDEO CODING Robert L. Stevenson Laboratory for Image and Signal Processing Department of Electrical Engineering University of Notre Dame Notre Dame, IN 46556

More information

NEW HYBRID FILTERING TECHNIQUES FOR REMOVAL OF GAUSSIAN NOISE FROM MEDICAL IMAGES

NEW HYBRID FILTERING TECHNIQUES FOR REMOVAL OF GAUSSIAN NOISE FROM MEDICAL IMAGES NEW HYBRID FILTERING TECHNIQUES FOR REMOVAL OF GAUSSIAN NOISE FROM MEDICAL IMAGES Gnanambal Ilango 1 and R. Marudhachalam 2 1 Postgraduate and Research Department of Mathematics, Government Arts College

More information

Image Classification Using Wavelet Coefficients in Low-pass Bands

Image Classification Using Wavelet Coefficients in Low-pass Bands Proceedings of International Joint Conference on Neural Networks, Orlando, Florida, USA, August -7, 007 Image Classification Using Wavelet Coefficients in Low-pass Bands Weibao Zou, Member, IEEE, and Yan

More information

Using the Holey Brick Tree for Spatial Data. in General Purpose DBMSs. Northeastern University

Using the Holey Brick Tree for Spatial Data. in General Purpose DBMSs. Northeastern University Using the Holey Brick Tree for Spatial Data in General Purpose DBMSs Georgios Evangelidis Betty Salzberg College of Computer Science Northeastern University Boston, MA 02115-5096 1 Introduction There is

More information

A reversible data hiding based on adaptive prediction technique and histogram shifting

A reversible data hiding based on adaptive prediction technique and histogram shifting A reversible data hiding based on adaptive prediction technique and histogram shifting Rui Liu, Rongrong Ni, Yao Zhao Institute of Information Science Beijing Jiaotong University E-mail: rrni@bjtu.edu.cn

More information

Title. Author(s)Smolka, Bogdan. Issue Date Doc URL. Type. Note. File Information. Ranked-Based Vector Median Filter

Title. Author(s)Smolka, Bogdan. Issue Date Doc URL. Type. Note. File Information. Ranked-Based Vector Median Filter Title Ranked-Based Vector Median Filter Author(s)Smolka, Bogdan Proceedings : APSIPA ASC 2009 : Asia-Pacific Signal Citationand Conference: 254-257 Issue Date 2009-10-04 Doc URL http://hdl.handle.net/2115/39685

More information

Introduction Distributed-memory parallel computers dominate today's parallel computing arena. These machines, such as the Kendall Square KSR-, Intel P

Introduction Distributed-memory parallel computers dominate today's parallel computing arena. These machines, such as the Kendall Square KSR-, Intel P Performance Comparison of a Set of Periodic and Non-Periodic Tridiagonal Solvers on SP2 and Paragon Parallel Computers Xian-He Sun Stuti Moitra Department of Computer Science Scientic Applications Branch

More information

m Environment Output Activation 0.8 Output Activation Input Value

m Environment Output Activation 0.8 Output Activation Input Value Learning Sensory-Motor Cortical Mappings Without Training Mike Spratling Gillian Hayes Department of Articial Intelligence University of Edinburgh mikes@dai.ed.ac.uk gmh@dai.ed.ac.uk Abstract. This paper

More information

AN ANALYSIS OF ITERATIVE ALGORITHMS FOR IMAGE RECONSTRUCTION FROM SATELLITE EARTH REMOTE SENSING DATA Matthew H Willis Brigham Young University, MERS Laboratory 459 CB, Provo, UT 8462 8-378-4884, FAX:

More information

The principal components of natural images. Peter J.B. Hancock, Roland J. Baddeley and Leslie S. Smith. University of Stirling, Scotland FK9 4LA

The principal components of natural images. Peter J.B. Hancock, Roland J. Baddeley and Leslie S. Smith. University of Stirling, Scotland FK9 4LA The principal components of natural images Peter J.B. Hancock, Roland J. Baddeley and Leslie S. Smith Centre for Cognitive and Computational Neuroscience Departments of Psychology and Computing Science

More information

Improving the Discrimination Capability with an Adaptive Synthetic Discriminant Function Filter

Improving the Discrimination Capability with an Adaptive Synthetic Discriminant Function Filter Improving the Discrimination Capability with an Adaptive Synthetic Discriminant Function Filter 83 J. Ángel González-Fraga 1, Víctor H. Díaz-Ramírez 1, Vitaly Kober 1, and Josué Álvarez-Borrego 2 1 Department

More information

Digital Image Processing. Prof. P. K. Biswas. Department of Electronic & Electrical Communication Engineering

Digital Image Processing. Prof. P. K. Biswas. Department of Electronic & Electrical Communication Engineering Digital Image Processing Prof. P. K. Biswas Department of Electronic & Electrical Communication Engineering Indian Institute of Technology, Kharagpur Lecture - 21 Image Enhancement Frequency Domain Processing

More information

Parallel Implementation of a Unied Approach to. Image Focus and Defocus Analysis on the Parallel Virtual Machine

Parallel Implementation of a Unied Approach to. Image Focus and Defocus Analysis on the Parallel Virtual Machine Parallel Implementation of a Unied Approach to Image Focus and Defocus Analysis on the Parallel Virtual Machine Yen-Fu Liu, Nai-Wei Lo, Murali Subbarao, Bradley S. Carlson yiu@sbee.sunysb.edu, naiwei@sbee.sunysb.edu

More information

INTERNATIONAL JOURNAL OF ELECTRONICS AND COMMUNICATION ENGINEERING & TECHNOLOGY (IJECET)

INTERNATIONAL JOURNAL OF ELECTRONICS AND COMMUNICATION ENGINEERING & TECHNOLOGY (IJECET) INTERNATIONAL JOURNAL OF ELECTRONICS AND COMMUNICATION ENGINEERING & TECHNOLOGY (IJECET) ISSN 0976 6464(Print) ISSN 0976 6472(Online) Volume 3, Issue 3, October- December (2012), pp. 153-161 IAEME: www.iaeme.com/ijecet.asp

More information

MULTICHANNEL image processing is studied in this

MULTICHANNEL image processing is studied in this 186 IEEE SIGNAL PROCESSING LETTERS, VOL. 6, NO. 7, JULY 1999 Vector Median-Rational Hybrid Filters for Multichannel Image Processing Lazhar Khriji and Moncef Gabbouj, Senior Member, IEEE Abstract In this

More information

Structural Similarity Optimized Wiener Filter: A Way to Fight Image Noise

Structural Similarity Optimized Wiener Filter: A Way to Fight Image Noise Structural Similarity Optimized Wiener Filter: A Way to Fight Image Noise Mahmud Hasan and Mahmoud R. El-Sakka (B) Department of Computer Science, University of Western Ontario, London, ON, Canada {mhasan62,melsakka}@uwo.ca

More information

Fast and Effective Interpolation Using Median Filter

Fast and Effective Interpolation Using Median Filter Fast and Effective Interpolation Using Median Filter Jian Zhang 1, *, Siwei Ma 2, Yongbing Zhang 1, and Debin Zhao 1 1 Department of Computer Science, Harbin Institute of Technology, Harbin 150001, P.R.

More information

Automatic Parameter Optimization for De-noising MR Data

Automatic Parameter Optimization for De-noising MR Data Automatic Parameter Optimization for De-noising MR Data Joaquín Castellanos 1, Karl Rohr 2, Thomas Tolxdorff 3, and Gudrun Wagenknecht 1 1 Central Institute for Electronics, Research Center Jülich, Germany

More information

New structural similarity measure for image comparison

New structural similarity measure for image comparison University of Wollongong Research Online Faculty of Engineering and Information Sciences - Papers: Part A Faculty of Engineering and Information Sciences 2012 New structural similarity measure for image

More information

2. CNeT Architecture and Learning 2.1. Architecture The Competitive Neural Tree has a structured architecture. A hierarchy of identical nodes form an

2. CNeT Architecture and Learning 2.1. Architecture The Competitive Neural Tree has a structured architecture. A hierarchy of identical nodes form an Competitive Neural Trees for Vector Quantization Sven Behnke and Nicolaos B. Karayiannis Department of Mathematics Department of Electrical and Computer Science and Computer Engineering Martin-Luther-University

More information

Wevelet Neuron Filter with the Local Statistics. Oriented to the Pre-processor for the Image Signals

Wevelet Neuron Filter with the Local Statistics. Oriented to the Pre-processor for the Image Signals Wevelet Neuron Filter with the Local Statistics Oriented to the Pre-processor for the Image Signals Noriaki Suetake Naoki Yamauchi 3 Takeshi Yamakawa y epartment of Control Engineering and Science Kyushu

More information

Weichuan Yu, Kostas Daniilidis, Gerald Sommer. Christian Albrechts University. Preusserstrasse 1-9, D Kiel, Germany

Weichuan Yu, Kostas Daniilidis, Gerald Sommer. Christian Albrechts University. Preusserstrasse 1-9, D Kiel, Germany Rotated Wedge Averaging Method for Junction Characterization Weichuan Yu, Kostas Daniilidis, Gerald Sommer Institute of Computer Science Christian Albrechts University Preusserstrasse -9, D-45 Kiel, Germany

More information

N.Priya. Keywords Compass mask, Threshold, Morphological Operators, Statistical Measures, Text extraction

N.Priya. Keywords Compass mask, Threshold, Morphological Operators, Statistical Measures, Text extraction Volume, Issue 8, August ISSN: 77 8X International Journal of Advanced Research in Computer Science and Software Engineering Research Paper Available online at: www.ijarcsse.com A Combined Edge-Based Text

More information

However, m pq is just an approximation of M pq. As it was pointed out by Lin [2], more precise approximation can be obtained by exact integration of t

However, m pq is just an approximation of M pq. As it was pointed out by Lin [2], more precise approximation can be obtained by exact integration of t FAST CALCULATION OF GEOMETRIC MOMENTS OF BINARY IMAGES Jan Flusser Institute of Information Theory and Automation Academy of Sciences of the Czech Republic Pod vodarenskou vez 4, 82 08 Prague 8, Czech

More information

Convolutional Neural Networks for Object Classication in CUDA

Convolutional Neural Networks for Object Classication in CUDA Convolutional Neural Networks for Object Classication in CUDA Alex Krizhevsky (kriz@cs.toronto.edu) April 16, 2009 1 Introduction Here I will present my implementation of a simple convolutional neural

More information

Low Contrast Image Enhancement Using Adaptive Filter and DWT: A Literature Review

Low Contrast Image Enhancement Using Adaptive Filter and DWT: A Literature Review Low Contrast Image Enhancement Using Adaptive Filter and DWT: A Literature Review AARTI PAREYANI Department of Electronics and Communication Engineering Jabalpur Engineering College, Jabalpur (M.P.), India

More information

Minoru SASAKI and Kenji KITA. Department of Information Science & Intelligent Systems. Faculty of Engineering, Tokushima University

Minoru SASAKI and Kenji KITA. Department of Information Science & Intelligent Systems. Faculty of Engineering, Tokushima University Information Retrieval System Using Concept Projection Based on PDDP algorithm Minoru SASAKI and Kenji KITA Department of Information Science & Intelligent Systems Faculty of Engineering, Tokushima University

More information

Feedback Guided Scheduling of Nested Loops

Feedback Guided Scheduling of Nested Loops Feedback Guided Scheduling of Nested Loops T. L. Freeman 1, D. J. Hancock 1, J. M. Bull 2, and R. W. Ford 1 1 Centre for Novel Computing, University of Manchester, Manchester, M13 9PL, U.K. 2 Edinburgh

More information

Using the Deformable Part Model with Autoencoded Feature Descriptors for Object Detection

Using the Deformable Part Model with Autoencoded Feature Descriptors for Object Detection Using the Deformable Part Model with Autoencoded Feature Descriptors for Object Detection Hyunghoon Cho and David Wu December 10, 2010 1 Introduction Given its performance in recent years' PASCAL Visual

More information

Image Denoising Based on Hybrid Fourier and Neighborhood Wavelet Coefficients Jun Cheng, Songli Lei

Image Denoising Based on Hybrid Fourier and Neighborhood Wavelet Coefficients Jun Cheng, Songli Lei Image Denoising Based on Hybrid Fourier and Neighborhood Wavelet Coefficients Jun Cheng, Songli Lei College of Physical and Information Science, Hunan Normal University, Changsha, China Hunan Art Professional

More information

Texture Image Segmentation using FCM

Texture Image Segmentation using FCM Proceedings of 2012 4th International Conference on Machine Learning and Computing IPCSIT vol. 25 (2012) (2012) IACSIT Press, Singapore Texture Image Segmentation using FCM Kanchan S. Deshmukh + M.G.M

More information

Structural Similarity Based Image Quality Assessment Using Full Reference Method

Structural Similarity Based Image Quality Assessment Using Full Reference Method From the SelectedWorks of Innovative Research Publications IRP India Spring April 1, 2015 Structural Similarity Based Image Quality Assessment Using Full Reference Method Innovative Research Publications,

More information

Face Hallucination Based on Eigentransformation Learning

Face Hallucination Based on Eigentransformation Learning Advanced Science and Technology etters, pp.32-37 http://dx.doi.org/10.14257/astl.2016. Face allucination Based on Eigentransformation earning Guohua Zou School of software, East China University of Technology,

More information

A Joint Histogram - 2D Correlation Measure for Incomplete Image Similarity

A Joint Histogram - 2D Correlation Measure for Incomplete Image Similarity A Joint Histogram - 2D Correlation for Incomplete Image Similarity Nisreen Ryadh Hamza MSc Candidate, Faculty of Computer Science & Mathematics, University of Kufa, Iraq. ORCID: 0000-0003-0607-3529 Hind

More information

PRINCIPAL COMPONENT ANALYSIS IMAGE DENOISING USING LOCAL PIXEL GROUPING

PRINCIPAL COMPONENT ANALYSIS IMAGE DENOISING USING LOCAL PIXEL GROUPING PRINCIPAL COMPONENT ANALYSIS IMAGE DENOISING USING LOCAL PIXEL GROUPING Divesh Kumar 1 and Dheeraj Kalra 2 1 Department of Electronics & Communication Engineering, IET, GLA University, Mathura 2 Department

More information

ADVANCED IMAGE PROCESSING METHODS FOR ULTRASONIC NDE RESEARCH C. H. Chen, University of Massachusetts Dartmouth, N.

ADVANCED IMAGE PROCESSING METHODS FOR ULTRASONIC NDE RESEARCH C. H. Chen, University of Massachusetts Dartmouth, N. ADVANCED IMAGE PROCESSING METHODS FOR ULTRASONIC NDE RESEARCH C. H. Chen, University of Massachusetts Dartmouth, N. Dartmouth, MA USA Abstract: The significant progress in ultrasonic NDE systems has now

More information

DIGITAL IMAGE WATERMARKING BASED ON A RELATION BETWEEN SPATIAL AND FREQUENCY DOMAINS

DIGITAL IMAGE WATERMARKING BASED ON A RELATION BETWEEN SPATIAL AND FREQUENCY DOMAINS DIGITAL IMAGE WATERMARKING BASED ON A RELATION BETWEEN SPATIAL AND FREQUENCY DOMAINS Murat Furat Mustafa Oral e-mail: mfurat@cu.edu.tr e-mail: moral@mku.edu.tr Cukurova University, Faculty of Engineering,

More information

then present results of directly applying the SFFS feature selection algorithms for mammographic mass detection, discussing how noise and the sie of t

then present results of directly applying the SFFS feature selection algorithms for mammographic mass detection, discussing how noise and the sie of t The role of feature selection in building pattern recogniers for computer-aided diagnosis Clay Spence and Paul Sajda National Information Display Laboratory Sarno Corporation Princeton, NJ 08543-5300,

More information

Assignment 3: Edge Detection

Assignment 3: Edge Detection Assignment 3: Edge Detection - EE Affiliate I. INTRODUCTION This assignment looks at different techniques of detecting edges in an image. Edge detection is a fundamental tool in computer vision to analyse

More information

Pattern based Residual Coding for H.264 Encoder *

Pattern based Residual Coding for H.264 Encoder * Pattern based Residual Coding for H.264 Encoder * Manoranjan Paul and Manzur Murshed Gippsland School of Information Technology, Monash University, Churchill, Vic-3842, Australia E-mail: {Manoranjan.paul,

More information

Robust Image Contour Detection by Watershed. Tampere University of Technology. P.O. Box 553, FIN Tampere, Finland

Robust Image Contour Detection by Watershed. Tampere University of Technology. P.O. Box 553, FIN Tampere, Finland 1 Robust Image Contour Detection by Watershed Transformation iao Pei and Moncef Gabbouj Signal Processing Laboratory Tampere University of Technology P.O. Box 553, FIN33101 Tampere, Finland Email: xiao@alpha.cc.tut.

More information

A ROBUST LONE DIAGONAL SORTING ALGORITHM FOR DENOISING OF IMAGES WITH SALT AND PEPPER NOISE

A ROBUST LONE DIAGONAL SORTING ALGORITHM FOR DENOISING OF IMAGES WITH SALT AND PEPPER NOISE International Journal of Computational Intelligence & Telecommunication Systems, 2(1), 2011, pp. 33-38 A ROBUST LONE DIAGONAL SORTING ALGORITHM FOR DENOISING OF IMAGES WITH SALT AND PEPPER NOISE Rajamani.

More information

LOCALIZATION OF FACIAL REGIONS AND FEATURES IN COLOR IMAGES. Karin Sobottka Ioannis Pitas

LOCALIZATION OF FACIAL REGIONS AND FEATURES IN COLOR IMAGES. Karin Sobottka Ioannis Pitas LOCALIZATION OF FACIAL REGIONS AND FEATURES IN COLOR IMAGES Karin Sobottka Ioannis Pitas Department of Informatics, University of Thessaloniki 540 06, Greece e-mail:fsobottka, pitasg@zeus.csd.auth.gr Index

More information

Keywords: Thresholding, Morphological operations, Image filtering, Adaptive histogram equalization, Ceramic tile.

Keywords: Thresholding, Morphological operations, Image filtering, Adaptive histogram equalization, Ceramic tile. Volume 3, Issue 7, July 2013 ISSN: 2277 128X International Journal of Advanced Research in Computer Science and Software Engineering Research Paper Available online at: www.ijarcsse.com Blobs and Cracks

More information

Reduction of Blocking artifacts in Compressed Medical Images

Reduction of Blocking artifacts in Compressed Medical Images ISSN 1746-7659, England, UK Journal of Information and Computing Science Vol. 8, No. 2, 2013, pp. 096-102 Reduction of Blocking artifacts in Compressed Medical Images Jagroop Singh 1, Sukhwinder Singh

More information

IMPLEMENTATION OF THE CONTRAST ENHANCEMENT AND WEIGHTED GUIDED IMAGE FILTERING ALGORITHM FOR EDGE PRESERVATION FOR BETTER PERCEPTION

IMPLEMENTATION OF THE CONTRAST ENHANCEMENT AND WEIGHTED GUIDED IMAGE FILTERING ALGORITHM FOR EDGE PRESERVATION FOR BETTER PERCEPTION IMPLEMENTATION OF THE CONTRAST ENHANCEMENT AND WEIGHTED GUIDED IMAGE FILTERING ALGORITHM FOR EDGE PRESERVATION FOR BETTER PERCEPTION Chiruvella Suresh Assistant professor, Department of Electronics & Communication

More information

A MORPHOLOGY-BASED FILTER STRUCTURE FOR EDGE-ENHANCING SMOOTHING

A MORPHOLOGY-BASED FILTER STRUCTURE FOR EDGE-ENHANCING SMOOTHING Proceedings of the 1994 IEEE International Conference on Image Processing (ICIP-94), pp. 530-534. (Austin, Texas, 13-16 November 1994.) A MORPHOLOGY-BASED FILTER STRUCTURE FOR EDGE-ENHANCING SMOOTHING

More information

Image Sampling and Quantisation

Image Sampling and Quantisation Image Sampling and Quantisation Introduction to Signal and Image Processing Prof. Dr. Philippe Cattin MIAC, University of Basel 1 of 46 22.02.2016 09:17 Contents Contents 1 Motivation 2 Sampling Introduction

More information

Using Local Trajectory Optimizers To Speed Up Global. Christopher G. Atkeson. Department of Brain and Cognitive Sciences and

Using Local Trajectory Optimizers To Speed Up Global. Christopher G. Atkeson. Department of Brain and Cognitive Sciences and Using Local Trajectory Optimizers To Speed Up Global Optimization In Dynamic Programming Christopher G. Atkeson Department of Brain and Cognitive Sciences and the Articial Intelligence Laboratory Massachusetts

More information

Value-and-Criterion Filters: A new filter structure based upon morphological opening and closing

Value-and-Criterion Filters: A new filter structure based upon morphological opening and closing Value-and-Criterion Filters: A new filter structure based upon morphological opening and closing Mark A. Schulze and John A. Pearce Department of Electrical and Computer Engineering and Biomedical Engineering

More information

Fuzzy Center Weighted Hybrid Filtering Techniques for Denoising of Medical Images

Fuzzy Center Weighted Hybrid Filtering Techniques for Denoising of Medical Images International Journal of Fuzzy Mathematics and Systems. ISSN 2248-9940 Volume 2, Number 4 (2012), pp. 383-390 Research India Publications http://www.ripublication.com Fuzzy Center Weighted Hybrid Filtering

More information

Image Sampling & Quantisation

Image Sampling & Quantisation Image Sampling & Quantisation Biomedical Image Analysis Prof. Dr. Philippe Cattin MIAC, University of Basel Contents 1 Motivation 2 Sampling Introduction and Motivation Sampling Example Quantisation Example

More information

OPTIMIZATION, OPTIMAL DESIGN AND DE NOVO PROGRAMMING: DISCUSSION NOTES

OPTIMIZATION, OPTIMAL DESIGN AND DE NOVO PROGRAMMING: DISCUSSION NOTES OPTIMIZATION, OPTIMAL DESIGN AND DE NOVO PROGRAMMING: DISCUSSION NOTES MILAN ZELENY Introduction Fordham University, New York, USA mzeleny@fordham.edu Many older texts, with titles like Globally Optimal

More information

Parallel Clustering on a Unidirectional Ring. Gunter Rudolph 1. University of Dortmund, Department of Computer Science, LS XI, D{44221 Dortmund

Parallel Clustering on a Unidirectional Ring. Gunter Rudolph 1. University of Dortmund, Department of Computer Science, LS XI, D{44221 Dortmund Parallel Clustering on a Unidirectional Ring Gunter Rudolph 1 University of Dortmund, Department of Computer Science, LS XI, D{44221 Dortmund 1. Introduction Abstract. In this paper a parallel version

More information

Transactions on Information and Communications Technologies vol 3, 1993 WIT Press, ISSN

Transactions on Information and Communications Technologies vol 3, 1993 WIT Press,   ISSN The implementation of a general purpose FORTRAN harness for an arbitrary network of transputers for computational fluid dynamics J. Mushtaq, A.J. Davies D.J. Morgan ABSTRACT Many Computational Fluid Dynamics

More information

Efficient Image Compression of Medical Images Using the Wavelet Transform and Fuzzy c-means Clustering on Regions of Interest.

Efficient Image Compression of Medical Images Using the Wavelet Transform and Fuzzy c-means Clustering on Regions of Interest. Efficient Image Compression of Medical Images Using the Wavelet Transform and Fuzzy c-means Clustering on Regions of Interest. D.A. Karras, S.A. Karkanis and D. E. Maroulis University of Piraeus, Dept.

More information

Fingerprint Classification Using Orientation Field Flow Curves

Fingerprint Classification Using Orientation Field Flow Curves Fingerprint Classification Using Orientation Field Flow Curves Sarat C. Dass Michigan State University sdass@msu.edu Anil K. Jain Michigan State University ain@msu.edu Abstract Manual fingerprint classification

More information

Coarse-to-Fine Search Technique to Detect Circles in Images

Coarse-to-Fine Search Technique to Detect Circles in Images Int J Adv Manuf Technol (1999) 15:96 102 1999 Springer-Verlag London Limited Coarse-to-Fine Search Technique to Detect Circles in Images M. Atiquzzaman Department of Electrical and Computer Engineering,

More information

Flexible-Hybrid Sequential Floating Search in Statistical Feature Selection

Flexible-Hybrid Sequential Floating Search in Statistical Feature Selection Flexible-Hybrid Sequential Floating Search in Statistical Feature Selection Petr Somol 1,2, Jana Novovičová 1,2, and Pavel Pudil 2,1 1 Dept. of Pattern Recognition, Institute of Information Theory and

More information

Neuro-Remodeling via Backpropagation of Utility. ABSTRACT Backpropagation of utility is one of the many methods for neuro-control.

Neuro-Remodeling via Backpropagation of Utility. ABSTRACT Backpropagation of utility is one of the many methods for neuro-control. Neuro-Remodeling via Backpropagation of Utility K. Wendy Tang and Girish Pingle 1 Department of Electrical Engineering SUNY at Stony Brook, Stony Brook, NY 11794-2350. ABSTRACT Backpropagation of utility

More information

Image Enhancement Techniques for Fingerprint Identification

Image Enhancement Techniques for Fingerprint Identification March 2013 1 Image Enhancement Techniques for Fingerprint Identification Pankaj Deshmukh, Siraj Pathan, Riyaz Pathan Abstract The aim of this paper is to propose a new method in fingerprint enhancement

More information

CHAPTER 6 PERCEPTUAL ORGANIZATION BASED ON TEMPORAL DYNAMICS

CHAPTER 6 PERCEPTUAL ORGANIZATION BASED ON TEMPORAL DYNAMICS CHAPTER 6 PERCEPTUAL ORGANIZATION BASED ON TEMPORAL DYNAMICS This chapter presents a computational model for perceptual organization. A figure-ground segregation network is proposed based on a novel boundary

More information

Image-Space-Parallel Direct Volume Rendering on a Cluster of PCs

Image-Space-Parallel Direct Volume Rendering on a Cluster of PCs Image-Space-Parallel Direct Volume Rendering on a Cluster of PCs B. Barla Cambazoglu and Cevdet Aykanat Bilkent University, Department of Computer Engineering, 06800, Ankara, Turkey {berkant,aykanat}@cs.bilkent.edu.tr

More information

A novel firing rule for training Kohonen selforganising

A novel firing rule for training Kohonen selforganising A novel firing rule for training Kohonen selforganising maps D. T. Pham & A. B. Chan Manufacturing Engineering Centre, School of Engineering, University of Wales Cardiff, P.O. Box 688, Queen's Buildings,

More information

Parallel Evaluation of Hopfield Neural Networks

Parallel Evaluation of Hopfield Neural Networks Parallel Evaluation of Hopfield Neural Networks Antoine Eiche, Daniel Chillet, Sebastien Pillement and Olivier Sentieys University of Rennes I / IRISA / INRIA 6 rue de Kerampont, BP 818 2232 LANNION,FRANCE

More information

Telecommunication and Informatics University of North Carolina, Technical University of Gdansk Charlotte, NC 28223, USA

Telecommunication and Informatics University of North Carolina, Technical University of Gdansk Charlotte, NC 28223, USA A Decoder-based Evolutionary Algorithm for Constrained Parameter Optimization Problems S lawomir Kozie l 1 and Zbigniew Michalewicz 2 1 Department of Electronics, 2 Department of Computer Science, Telecommunication

More information