SEARCH · PubMed Health

Results for “Parallel Algorithms”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 343 records · Page 19Linked to original sources

PubMedDate not supplied

Retrospective registration of PET and MR brain images: an algorithm and its stereotactic validation.

OBJECTIVE: We present a validation study of an algorithm for retrospective registration of PET and MR brain images. MATERIALS AND METHODS: This algorithm involves two steps. In the first step, the two volumes are reformatted by aligning their interhemispheric fissure planes (midsagittal plane). In the second step, the corresponding planes parallel to the midsagittal plane are further aligned in the reformatted volumes to produce a 3D rigid body registration of the two original volumes. It is an efficient algorithm because both steps are performed in 2D spaces, and in each step only a small number of landmarks are required. A user-friendly system has been implemented to facilitate easy and fast processing of registration and reformatting of image volumes. The accuracy of this algorithm is validated using clinical scans of neurosurgical patients with a stereotaxic frame attached to their skull. The frame-based stereotaxic system provides an effective method for transforming image coordinates from different image volumes into a common coordinate system. This common coordinate system is used for assessing the spatial correspondence of each pixel in the registered image volumes. Validation using the stereotaxic image volumes enables objective estimation of retrospective registration accuracy. RESULTS: Analysis of 11 MR/PET image pairs indicates that our registration method not only is efficient but also provides adequate accuracy for most clinical evaluation of PET studies. CONCLUSION: We have implemented and validated an efficient algorithm for retrospective registration of PET and MR brain images.

Adult↗

PubMed2001

Efficiency of parallel direct optimization.

Tremendous progress has been made at the level of sequential computation in phylogenetics. However, little attention has been paid to parallel computation. Parallel computing is particularly suited to phylogenetics because of the many ways large computational problems can be broken into parts that can be analyzed concurrently. In this paper, we investigate the scaling factors and efficiency of random addition and tree refinement strategies using the direct optimization software, POY, on a small (10 slave processors) and a large (256 slave processors) cluster of networked PCs running LINUX. These algorithms were tested on several data sets composed of DNA and morphology ranging from 40 to 500 taxa. Various algorithms in POY show fundamentally different properties within and between clusters. All algorithms are efficient on the small cluster for the 40-taxon data set. On the large cluster, multibuilding exhibits excellent parallel efficiency, whereas parallel building is inefficient. These results are independent of data set size. Branch swapping in parallel shows excellent speed-up for 16 slave processors on the large cluster. However, there is no appreciable speed-up for branch swapping with the further addition of slave processors (>16). This result is independent of data set size. Ratcheting in parallel is efficient with the addition of up to 32 processors in the large cluster. This result is independent of data set size.

Algorithms↗

PubMed2002

Feature decomposition architectures for neural networks: algorithms, error bounds, and applications.

In recent years, systems consisting of multiple modular neural networks have attracted substantial interest in the neural networks community because of various advantages they offer over a single large monolithic network. In this paper, we propose two basic feature decomposition models (namely, parallel model and tandem model) in which each of the neural network modules processes a disjoint subset of the input features. A novel feature decomposition algorithm is introduced to partition the input space into disjoint subsets solely based on the available training data. Under certain assumptions, the approximation error due to decomposition can be proved to be bounded by any desired small value over a compact set. Finally, the performance of feature decomposition networks is compared with that of a monolithic network in real world bench mark pattern recognition and modeling problems.

Algorithms↗

PubMed1999

A study of deleterious gene structure in plants using Markov chain Monte Carlo.

The characteristics of deleterious genes have been of great interest in both theory and practice in genetics. Because of the complex genetic mechanism of these deleterious genes, most current studies try to estimate the overall magnitude of mortality effects on a population, which is characterized classically by the number of lethal equivalents. This number is a combination of several parameters, each of which has a distinct biological effect on genetic mortality. In conservation and breeding programs, it is important to be able to distinguish among different combinations of these parameters that lead to the same number of lethal equivalents, such as a large number of mildly deleterious genes or a few lethal genes, The ability to distinguish such parameter combinations requires more than one generation of mating. We propose a model for survival data from a two-generation mating experiment on the plant species Brassica rapa, and we enable inference with Markov chain Monte Carlo. This computational strategy is effective because a vast amount of missing genotype information must be accounted for. In addition to the lethal equivalents, the two-generation data provide separate information on the average intensity of mortality and the average number of deleterious genes carried by an individual. In our Markov chain Monte Carlo algorithm, we use a vector proposal distribution to overcome inefficiency of a single-site Gibbs sampler. Information about environmental effects is obtained from an outcrossing experiment conducted in parallel with the two-generation mating experiments.

Algorithms↗

PubMed1998

Detection of suspected malignant patterns in three-dimensional magnetic resonance breast images.

In this article, a Boolean Neural Network (BNN) is used for the detection of suspected malignant regions in 3D breast magnetic resonance (MR) images. The BNN is characterized by fast learning and classification, guaranteed convergence, and simple, integer weight calculations. The BNN learning algorithm is incremental, which allows the addition and deletion of training patterns without unlearning those already learned. The incremental learning algorithm automatically reduces the training set and trains the network only with those examples estimated to be useful. The architecture is suitable for parallel hardware implementation using available Very Large Scale Integration (VLSI) technology. The BNN was trained by using a set of malignant, benign, and false-positive patterns, extracted by experts, from selected MR studies, by using an incremental learning algorithm. After training, the network was tested by means of a consistency checking test, cross validation techniques, and patterns from actual MR breast images. During the consistency test, the BNN was tested by using the same patterns used for training. The BNN classification accuracy in this case was 99.75%, proving the ability of the BNN to select useful patterns from the training set. Then, a leave one out cross-validation (LOOCV) test was done by using patterns from the training set and the classification accuracy was 90%. Next, an extended training set was created by shifting the original patterns in different directions. A cross-validation test was then performed by dividing the set of patterns into a training and a test set. Classification accuracy was compared to the nearest neighbor classifier. Results showed that the BNN achieved an average of 77% classification accuracy while requiring only 34% of the original training set. On the other hand, the nearest neighbor classifier achieved an accuracy of 57.9% while retaining the whole training set. Another test using actual MR slices different from the training set was done and results compared favorably to a radiologist's findings. Test results show the BNN's capability to detect suspected malignant regions in 3D MR images of the breast. The proposed BNN architecture can save the radiologist a great deal of time browsing MR slices searching for suspected malignancies.

Algorithms↗

PubMed1993

Faster sequential genetic linkage computations.

Linkage analysis using maximum-likelihood estimation is a powerful tool for locating genes. As available data sets have grown, the computation required for analysis has grown exponentially and become a significant impediment. Others have previously shown that parallel computation is applicable to linkage analysis and can yield order-of-magnitude improvements in speed. In this paper, we demonstrate that algorithmic modifications can also yield order-of-magnitude improvements, and sometimes much more. Using the software package LINKAGE, we describe a variety of algorithmic improvements that we have implemented, demonstrating both how these techniques are applied and their power. Experiments show that these improvements speed up the programs by an order of magnitude, on problems of moderate and large size. All improvements were made only in the combinatorial part of the code, without restoring to parallel computers. These improvements synthesize biological principles with computer science techniques, to effectively restructure the time-consuming computations in genetic linkage analysis.

Algorithms↗

PubMed1997

Secondary structure computer prediction of the poliovirus 5' non-coding region is improved by a genetic algorithm.

Comparison of the secondary structure of the 5' non-coding region of poliovirus 3 RNA derived from the genetic algorithm with the model of Skinner et al. (J. Mol. Biol., 207, 379-392, 1989) demonstrates many of the confirmed structural elements. The genetic algorithm (Shapiro and Navetta, J. Supercomput., 8, 195-201, 1994) generates a population of all possible stems, then mixes, combines, and recombines these stems in multiple iterations on a massively parallel computer, ultimately selecting a most fit structure based on its energy. The secondary structure of the region containing the determinants of neurovirulence was better predicted using the genetic algorithm, whereas the dynamic programming algorithm (Zuker, Science, 244, 48-52, 1989) required phylogenetic comparative sequence analysis to arrive at the correct conclusion. In addition, artificial mutations were introduced throughout this region of the genome and although rearrangements in structure may occur, many structures persisted, suggesting that the given structures thus selected may have evolved to withstand isolated mutations. The genetic algorithm-derived structure for the 5' non-coding region compares favorably with the biological data and functions previously described, and contains all of the 'persistent' structures, suggesting also that the persistence factor may be an aid to validating structures.

Algorithms↗

PubMed1995

Supercomputer algorithms for efficient linear octree encoding of three-dimensional brain images.

We designed and implemented algorithms for three-dimensional (3-D) reconstruction of brain images from serial sections using two important supercomputer architectures, vector and parallel. These architectures were represented by the Cray YMP and Connection Machine CM-2, respectively. The programs operated on linear octree representations of the brain data sets, and achieved 500-800 times acceleration when compared with a conventional laboratory workstation. As the need for higher resolution data sets increases, supercomputer algorithms may offer a means of performing 3-D reconstruction well above current experimental limits.

Algorithms↗

PubMed1991

[Response analysis for an approximate 3-D image reconstruction in cone-beam SPECT].

Cone-beam Single Photon Emission Computed Tomography (SPECT) offers the potential for a large increase in sensitivity as compared with parallel hole or fan-beam collimation. Three-dimensional image reconstruction was approximately accomplished by backprojecting filtered projections using a two-dimensional fan-beam algorithm. The cone-beam projection data were formed from mathematical phantoms as analytically derived line integrals of the density. In order to reduce the processing time, the filtered projections were backprojected into each planes parallel to the circle on which the focal point moved. Discrepancy of source position and degradation of resolution were investigated by computer simulation in three-dimensional image space. The obtained results suggest that, the nearer to the central plane or the axis of rotation, the less image degradation is performed. By introducing a parameter of angular difference between the focal point and fixed point in the image space during rotation, degradation of the reconstructed image can be estimated for any cone-beam SPECT system.

Algorithms↗

PubMed2002

YIN, a fundamental frequency estimator for speech and music.

An algorithm is presented for the estimation of the fundamental frequency (F0) of speech or musical sounds. It is based on the well-known autocorrelation method with a number of modifications that combine to prevent errors. The algorithm has several desirable features. Error rates are about three times lower than the best competing methods, as evaluated over a database of speech recorded together with a laryngograph signal. There is no upper limit on the frequency search range, so the algorithm is suited for high-pitched voices and music. The algorithm is relatively simple and may be implemented efficiently and with low latency, and it involves few parameters that must be tuned. It is based on a signal model (periodic signal) that may be extended in several ways to handle various forms of aperiodicity that occur in particular applications. Finally, interesting parallels may be drawn with models of auditory processing.

Algorithms↗

PubMed1987

Optimization by stimulating molecular evolution.

Based on the analogy between mathematical optimization and molecular evolution and on Eigen's quasi-species model of molecular evolution, an evolutionary algorithm for combinatorial optimization has been developed. This algorithm consists of a versatile variation scheme and an innovative decision rule, the essence of which lies in a radical revision of the conventional philosophy of optimization: A number of configurations of variables with better values, instead of only a single best configuration, are selected as starting points for the next iteration. As a result the search proceeds in parallel along a number of routes and is unlikely to get trapped in local optima. An important innovation of the algorithm is introduction of a constraint to let the starting points always keep a certain distance from each other so that the search is able to cover a larger region of space effectively. The main advantage of the algorithm is that it has more chances to find the global optimum and as many local optima as possible in a single run. This has been demonstrated in preliminary computational experiments.

Algorithms↗

PubMed2000

A real-time volumetric visualization system for electrical impedance tomography.

Three dimensional (3D) electrical impedance tomography (EIT) presents many additional challenges over and above those associated with two dimensional EIT systems. With present two dimensional (2D) systems, tomographs can be reconstructed and displayed on a PC with a standard computer monitor. In addition, using appropriate data acquisition hardware and simple image reconstruction algorithms, it is possible to collect, reconstruct and display volumetric EIT images in real time using parallel processing architectures. The advantages of this 'real-time' capability are many and include the ability to immediately assess the correct functioning of the system and the ability to track patient events and the effect of procedures in real time. Whilst 3D EIT boundary datasets can be collected in real time, their real-time image reconstruction and display presents some computational challenges. This explains why, to date, no real-time solutions have been presented. In addition the use of a standard computer monitor to display 3D volumes is unsatisfactory since not all depth cues are preserved when using this type of 2D display device. We present a system which is capable of displaying 3D EIT datasets in real time and allows interactive modification of the user's viewpoint. This allows the user to fly around (and through) the EIT volumetric dataset.

Algorithms↗

PubMed1996

Automated particle classification based on digital acquisition and analysis of flow cytometric pulse waveforms.

In flow cytometry, the typical use of front-end analog processing limits the pulse waveform features that can be measured to pulse integral, height, and width. Direct digitizing of the waveforms provides a means for the extraction of additional features, for example, pulse skewness and kurtosis, and Fourier properties. In this work, we have first demonstrated that the Fourier properties of the pulse can be employed usefully for discrimination between different types of cells that otherwise cannot be classified by using only time-domain features of the pulse. We then implemented and evaluated automatic procedures for cell classification based on neural networks. We established that neural networks could provide an efficient means of classification of cell types without the need for user interaction. The neural networks were also employed in an innovative manner for analysis of the digital flow cytometric data without feature extraction. The performance of the neural networks was compared with that of a more conventional means of classification, the K-means clustering algorithm. Neural networks can be realized in hardware, and this, in addition to their highly parallel architecture, makes them an important potential part of real-time analysis systems. These results are discussed in terms of the design of a real-time digital data acquisition system for flow cytometry.

Animals↗

PubMed1993

Quantitative analysis of gel electrophoretograms by image analysis and least squares modeling.

A computer-aided quantitative method for a complex analysis of gel electrophoretograms is presented. The analysis consists of several steps: (i) determination of the background image by methods of mathematical morphology and its subtraction from the gel image, (ii) selection of an appropriate part of the gel lane including curved lanes and lanes with a nonuniform width, (iii) computation of the lane densitogram by averaging several lane-parallel scans, (iv) decomposition of the lane densitogram into component bands using a data selecting algorithm and Marquardt's minimizer. Several different functions for component bands are utilized. It is shown that the densitogram can be decomposed into component bands with reasonable accuracy only if an appropriate model function is chosen. The algorithms are tested on several different gel electrophoretograms which show typical features as a nonuniform background, curved lanes, an asymmetrical band shape and a superposition of small bands on the shoulders of big ones. It is shown that overlapped bands are best approximated by an asymmetrical Gausian curve and an asymmetrical Gauss-Cauchy function. Linear response to the serial dilution of the protein sample is tested.

DNA, Bacterial↗

PubMed2003

Molecular computing revisited: a Moore's Law?

Moore's Law states that the processing power of microchips doubles every one to two years. This observation might apply to the nascent field of molecular computing, in which biomolecules carry out logical operations. Incorporation of new technologies that improve sensitivity and throughput has increased the complexity of problems that can be addressed. It is an ultimate goal for molecular computers to use the full potential of massive parallelism.

Algorithms↗

PubMed1989

The clinical utility of renal concentrating capacity in polycystic kidney disease.

We studied 177 adult nonazotemic subjects with autosomal dominant polycystic kidney disease (ADPKD) and 123 unaffected family members (NADPKD). In order to assess the factors influencing renal concentrating capacity maximal urinary osmolality (UOsm) after overnight water deprivation and vasopressin was measured. UOsm was reduced in ADPKD (680 +/- 14) compared to NADPKD subjects (812 +/- 13 mOsm/kg). A greater severity of the architectural abnormality as assessed by cyst number and size and remaining volume of normal parenchyma was associated with a greater impairment of renal concentrating capacity. The concentrating defect was present in the youngest ADPKD subjects and the rate of decline of concentrating capacity with age in ADPKD paralleled that in NADPKD subjects. Based on the initial 135 subjects studied, we developed an algorithm for diagnostic screening for ADPKD utilizing blood pressure, serum creatinine and UOsm designed to maximize sensitivity. When applied to a subsequent population of 165 adults, 121 with ADPKD and 44 unaffected relatives, this algorithm would have spared 20% of unaffected subjects from the cost of ultrasound while failing to detect less than 2% of affected subjects. This simple protocol thus offers a rapid and inexpensive way to screen for ADPKD.

Adult↗

PubMed1982

Neural networks and physical systems with emergent collective computational abilities.

Computational properties of use of biological organisms or to the construction of computers can emerge as collective properties of systems having a large number of simple equivalent components (or neurons). The physical meaning of content-addressable memory is described by an appropriate phase space flow of the state of a system. A model of such a system is given, based on aspects of neurobiology but readily adapted to integrated circuits. The collective properties of this model produce a content-addressable memory which correctly yields an entire memory from any subpart of sufficient size. The algorithm for the time evolution of the state of the system is based on asynchronous parallel processing. Additional emergent collective properties include some capacity for generalization, familiarity recognition, categorization, error correction, and time sequence retention. The collective properties are only weakly sensitive to details of the modeling or the failure of individual devices.

Animals↗

PubMed2000

Reconstruction of two- and three-dimensional images from synthetic-collimator data.

A novel SPECT collimation method, termed the synthetic collimator, is proposed. The synthetic collimator employs a multiple-pinhole aperture and a high-resolution detector. The problem of multiplexing, normally associated with multiple pinholes, is reduced by obtaining projections at a number of pinhole-detector distances. Projections with little multiplexing are collected at small pinhole-detector distances and high-resolution projections are collected at greater pinhole-detector distances. These projections are then reconstructed using the ML-EM algorithm. It is demonstrated through computer simulations that the synthetic collimator has superior resolution properties to a high-resolution parallel-beam (HRPB) collimator and a specially built ultra-high-resolution parallel-beam (UHRPB) collimator designed for our 0.38-mm pixel CdZnTe detectors. It is also shown that reconstructing images in three dimensions is superior to reconstructing them in two dimensions. The advantages of a high-resolution synthetic collimator over the parallel-hole collimators are apparently reduced in the presence of statistical noise. However, a high-sensitivity synthetic collimator was designed which again shows superior properties to the parallel-hole collimators. Finally, it is demonstrated that, for the cases studied, high-resolution detectors are necessary for the proper functionality of the synthetic collimator.

Algorithms↗