SEARCH · PubMed Health

Results for “Gaussian process”

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 703 records · Page 39Linked to original sources

PubMed2006

Identification, segmentation, and image property study of acute infarcts in diffusion-weighted images by using a probabilistic neural network and adaptive Gaussian mixture model.

RATIONALE AND OBJECTIVES: Accurate identification of infarcted regions of the brain is critical in management of stroke patients. An efficient and fast method for identification and segmentation of infarcts in the diffusion-weighted images (DWI) is proposed. MATERIALS AND METHODS: Thirteen stroke patients were studied. DWI scans were acquired with a slice thickness of 5 mm. We have used a probabilistic neural network for selecting infarct slices and an adaptive (two-level) Gaussian mixture model for segmentation of the infarcts. Statistical analysis, such as identification of distribution, first-order statistics calculation, and receiver operating characteristic curve analysis, was performed. RESULTS: The average dice index is about 0.6, and average sensitivity and specificity are about 81% and 99%, respectively. The value of sensitivity and dice index are influenced by the number of false positives and false negatives. Because artifacts and infarcts have similar imaging characteristics, it is difficult to completely eliminate the artifacts. The accuracy of localization is nearly 100% as there were only two false-positive and three false-negative slices of all 381 slices. The algorithm takes about 1 minute in the Matlab computing environment to process a volume. CONCLUSION: A method to localize and segment the acute brain infarcts is proposed. The method aids the clinician in reducing the time needed to localize and segment the infarcts. The speed of localization and segmentation can be enhanced further by implementing the algorithm in VC++ and using fast algorithms for selection of Gaussian mixture model parameters.

Algorithms↗

PubMed2004

Measurement of retinal vessel widths from fundus images based on 2-D modeling.

Changes in retinal vessel diameter are an important sign of diseases such as hypertension, arteriosclerosis and diabetes mellitus. Obtaining precise measurements of vascular widths is a critical and demanding process in automated retinal image analysis as the typical vessel is only a few pixels wide. This paper presents an algorithm to measure the vessel diameter to subpixel accuracy. The diameter measurement is based on a two-dimensional difference of Gaussian model, which is optimized to fit a two-dimensional intensity vessel segment. The performance of the method is evaluated against Brinchmann-Hansen's half height, Gregson's rectangular profile and Zhou's Gaussian model. Results from 100 sample profiles show that the presented algorithm is over 30% more precise than the compared techniques and is accurate to a third of a pixel.

Algorithms↗

PubMed1989

Automatic detection method of stenotic lesions in coronary cineangiograms.

A new image processing procedure enabling the automatic detection of coronary artery stenoses by cineangiography was developed. Detection of stenoses was performed using computer image processing by the following procedure. The path of the arteries was extracted by a subtraction method. The thresholding image was obtained from the subtraction image, and converted to a 'thinning' image, which represented the center line of the artery. For measurement of the arterial diameter, the vessel edges were determined by unilateral Gaussian fit to profile curves in sections perpendicular to the center line. Stenoses could then be detected on the basis of the normal diameter of artery estimated by Hough transformation. This method facilitates the detection of stenotic lesions from coronary cineangiograms.

Cineangiography↗

PubMedDate not supplied

Stochastic resonance in noisy spiking retinal and sensory neuron models.

Two new theorems show that small amounts of additive white noise can improve the bit count or mutual information of several popular models of spiking retinal neurons and spiking sensory neurons. The first theorem gives necessary and sufficient conditions for this noise benefit or stochastic resonance (SR) effect for subthreshold signals in a standard family of Poisson spiking models of retinal neurons. The result holds for all types of finite-variance noise and for all types of infinite-variance stable noise: SR occurs if and only if a sum of noise means or location parameters falls outside a 'forbidden interval' of values. The second theorem gives a similar forbidden-interval sufficient condition for the SR effect for several types of spiking sensory neurons that include the Fitzhugh-Nagumo neuron, the leaky integrate-and-fire neuron, and the reduced Type I neuron model if the additive noise is Gaussian white noise. Simulations show that neither the forbidden-interval condition nor Gaussianity is necessary for the SR effect.

Algorithms↗

PubMed1998

A method for approximating the density of maximum-likelihood and maximum a posteriori estimates under a Gaussian noise model.

The performance of maximum-likelihood (ML) and maximum a posteriori (MAP) estimates in non-linear problems at low data SNR is not well predicted by the Cramér-Rao or other lower bounds on variance. In order to better characterize the distribution of ML and MAP estimates under these conditions, we derive a point approximation to density values of the conditional distribution of such estimates. In an example problem, this approximate distribution captures the essential features of the distribution of ML estimates in the presence of Gaussian-distributed noise.

Algorithms↗

PubMed2001

Blind source separation using temporal predictability.

A measure of temporal predictability is defined and used to separate linear mixtures of signals. Given any set of statistically independent source signals, it is conjectured here that a linear mixture of those signals has the following property: the temporal predictability of any signal mixture is less than (or equal to) that of any of its component source signals. It is shown that this property can be used to recover source signals from a set of linear mixtures of those signals by finding an un-mixing matrix that maximizes a measure of temporal predictability for each recovered signal. This matrix is obtained as the solution to a generalized eigenvalue problem; such problems have scaling characteristics of O(N3), where N is the number of signal mixtures. In contrast to independent component analysis, the temporal predictability method requires minimal assumptions regarding the probability density functions of source signals. It is demonstrated that the method can separate signal mixtures in which each mixture is a linear combination of source signals with supergaussian, subgaussian, and gaussian probability density functions and on mixtures of voices and music.

Humans↗

PubMed2002

A feature-segmentation model of short-term visual memory.

A feature-segmentation model of short-term visual memory (STVM) for contours is proposed. Memory of the First stimulus is maintained until the second stimulus is observed. Three processes interact to determine the relationship between stimulus and response: feature encoding, memory, and decision. Basic assumptions of the model are twofold: (i) the STVM system divides a contour into convex parts at regions of concavity; and (ii) the value of each convex part represented in STVM is an independent Gaussian random variable. Simulation showed that the five-parameter fits give a good account of the effects of the four experimental variables. The model provides evidence that: (i) contours are successfully encoded within 0.5 s exposure, regardless of pattern complexity; (ii) memory noise increases as a linear function of retention interval; (iii) the capacity of STVM, defined by pattern complexity (the degree that a pattern can be handled for several seconds with little loss), is about 4 convex parts; and (iv) the confusability contributing to the decision process is a primary factor in deteriorating recognition of complex figures. It is concluded that visually presented patterns can be retained in STVM with considerable precision for prolonged periods of time, though some loss of precision is inevitable.

Humans↗

PubMed2004

[Dynamic contour tracking of medical images based on improved particle filter].

In the research of medical image processing, motion estimation and tracking relating to the region of interest has been given considerable attention. For improving the quality of the noisy or cluttered medical images, the particle filter (PF) based on the non-linear and non-Gaussian Bayesian State Estimation is a better as well as a technically challenging solution. As the algorithm of particle weights, especially the importance density function, often severely affects the performance of the PF, we propose in this paper a better algorithm for its improvement; in addition, to ensure better tracking of the dynamic contour with the PF, we proposed a new algorithm for the likelihood and prior probability density. Objective theoretical evaluation and substantial comparative experiments suggest that this method can be a good solution for accurate dynamic contour tracking.

Algorithms↗

PubMed2000

Rapid and effective correction of RF inhomogeneity for high field magnetic resonance imaging.

The well-known variability in the distribution of high frequency electromagnetic fields in the human body causes problems in the analysis of structural information in high field magnetic resonance images. We describe a method of compensating for the purely intensity-based effects. In our simple and rapid correction algorithm, we first use statistical means to determine the background image noise level and the edges of the image features. We next populate all "noise" pixels with the mean signal intensity of the image features. These data are then smoothed by convolution with a gaussian filter using Fourier methods. Finally, the original data that are above the noise level are normalized to the smoothed images, thereby eliminating the lowest spatial frequencies in the final, corrected data. Processing of a 124 slice, 256 x 256 volume dataset requires under 70 sec on a laptop personal computer. Overall, the method is less prone to artifacts from edges or from sensitivity to absolute head position than are other correction techniques. Following intensity correction, the images demonstrated obvious qualitative improvement and, when subjected to automated segmentation tools, the accuracy of segmentation improved, in one example, from 35.3% to 84.7% correct, as compared to a manually-constructed gold standard.

Algorithms↗

PubMed1993

Separate populations of visually guided saccades in humans: reaction times and amplitudes.

The saccadic eye movements of 20 naive adults, 7 naive teenagers, 12 naive children, and 4 trained adult subjects were measured using two single target saccade tasks; the gap and the overlap task. In the gap task, the fixation point was switched off before the target occurred; in the overlap task it remained on until the end of each trial. The target position was randomly selected 4 degrees to the left or 4 degrees to the right of the fixation point. The subjects were instructed to look at the target when it appeared, not to react as fast as possible. They were not given any feedback about their performance. The results suggest that, in the gap task, most of the naive subjects exhibit at least two (the teenagers certainly three) clearly separated peaks in the distribution of the saccadic reaction times. The first peak occurs between 100 and 135 ms (express saccades), the second one between 140 and 180 ms (fast regular), and a third peak may follow at about 200 ms (slow regular). Other subjects did not show clear signs of two modes in the range of 100 to 180 ms, and still others did not produce any reaction times below 135 ms. In the overlap task as well three or even more peaks were obtained at about the same positions along the reaction time scale of many, but not all subjects. Group data as well as those of individual subjects were fitted by the superposition of three gaussian functions. Segregating the reaction time data into saccades that over- or undershoot the target indicated that express saccades almost never overshoot. The results are discussed in relation to the different neural processes preceding the initiation of visually-guided saccades.

Adolescent↗

PubMed2005

Applying spatial distribution analysis techniques to classification of 3D medical images.

OBJECTIVE: The objective of this paper is to classify 3D medical images by analyzing spatial distributions to model and characterize the arrangement of the regions of interest (ROIs) in 3D space. METHODS AND MATERIAL: Two methods are proposed for facilitating such classification. The first method uses measures of similarity, such as the Mahalanobis distance and the Kullback-Leibler (KL) divergence, to compute the difference between spatial probability distributions of ROIs in an image of a new subject and each of the considered classes represented by historical data (e.g., normal versus disease class). A new subject is predicted to belong to the class corresponding to the most similar dataset. The second method employs the maximum likelihood (ML) principle to predict the class that most likely produced the dataset of the new subject. RESULTS: The proposed methods have been experimentally evaluated on three datasets: synthetic data (mixtures of Gaussian distributions), realistic lesion-deficit data (generated by a simulator conforming to a clinical study), and functional MRI activation data obtained from a study designed to explore neuroanatomical correlates of semantic processing in Alzheimer's disease (AD). CONCLUSION: Performed experiments demonstrated that the approaches based on the KL divergence and the ML method provide superior accuracy compared to the Mahalanobis distance. The later technique could still be a method of choice when the distributions differ significantly, since it is faster and less complex. The obtained classification accuracy with errors smaller than 1% supports that useful diagnosis assistance could be achieved assuming sufficiently informative historic data and sufficient information on the new subject.

Algorithms↗

PubMed2006

Dynamics of bolaamphiphilic fluorescent polyenes in lipid bilayers from polarization emission spectroscopy.

The rotational motions of the biamphiphilic polyenes (bolapolyenes) dimethyl all-(E)-octacosa-10,12,14,16,18-pentaenedioate (DE28:5) and dimethyl all-(E)-tetratriaconta-13,15,17,19,21-pentaenedioate (DE34:5), with head-to-head distances of 34 and 42A, respectively, have been examined by fluorescence anisotropy methods. The membrane-spanning bolapolyenes, which contain a central emitting pentaene group tethered to two methoxycarbonyl opposite polar heads by symmetric C(8) (DE28:5) and C(11) (DE34:5) polymethylene chains, were dispersed in lipid bilayers of DPPC or DMPC, and the stationary and picosecond-resolved emission was recorded as a function of temperature. In fluid-phase DMPC bilayers, three relaxation times could be determined, assigned to fast (0.2 and 2ns) single-bond isomerization processes localized on the alkyl chains, and to whole-molecule oscillations ( approximately 11ns), respectively. The anisotropy decay parameters were further analyzed in terms of a diffusive model for wobbling in a Gaussian ordering potential, to assess the anchoring effect of the symmetric polar heads. In this way, the average rotational diffusion constant of the bolapolyenes, D( perpendicular), could be estimated as 0.022-0.026rad(2) ns(-1) (DMPC bilayers, 35 degrees Celsius), a value that is only 1/3 of that corresponding to the related pentaene fatty acid spanning a single membrane monolayer. In contrast, the amplitude of the equilibrium orientational distribution (theta(half-cone) approximately 50 degrees ) is very similar for both the transmembrane and the single-headed polyenes. The reorientational oscillations of the central emitting group in the bolapolyenes necessarily would produce large-amplitude (2-5A) and very fast (ns) translational motions of the polar heads.

Lipid Bilayers↗

PubMed2004

In situ metrology of 85Kr plumes released by the COGEMA La Hague nuclear reprocessing plant.

The IRSN (Institut de Radioprotection et de Sûreté Nucléaire, France) has started an in situ study of the behaviour of atmospheric releases close-by the COGEMA La Hague nuclear reprocessing plant. The study is designed to improve information on the dispersion of radioactive pollutants very close to the emission point--a 100 m height chimney. In this situation, close to the emission and height of the emission, Gaussian models generally used to predict the behaviour of atmospheric releases are not well adapted. The study is based on the characterisation of the 85Kr emitted during normal operations of the reprocessing process. Temporal and spatial variations of the plume shape were investigated with intensive in situ measurements. Live in situ techniques to measure the electrons and the photons emitted by the 85Kr have been implemented and will be described. Preliminary results showing the interest of the techniques were presented. Variations of the dose rates created by the photon flux of a plume were directly measured and correlated to other quantities.

Belgium↗

PubMed2002

Diffusion entropy and waiting time statistics of hard-x-ray solar flares.

We show at work a technique of scaling detection based on evaluating the Shannon entropy of the diffusion process obtained by converting the time series under study into trajectories. This method, called diffusion entropy, affords information that cannot be derived from the direct evaluation of waiting times. We apply this method to the analysis of the distribution of time distance tau between two nearest-neighbor solar flares. This traditional part of the analysis is based on the direct evaluation of the distribution function psi(tau), or of the probability Psi(tau), that no time distance smaller than a given tau is found. We adopt the paradigm of the inverse power-law behavior, and we focus on the determination of the inverse power index mu, without ruling out different asymptotic properties that might be revealed, at larger scales, with the help of richer statistics. We then use the DE method, with three different walking rules, and we focus on the regime of transition to scaling. This regime of transition and the value of the scaling parameter itself, delta, depends on the walking rule adopted, a property of interest to shed light on the slow process of transition from dynamics to thermodynamics often occurring under anomalous statistical conditions. With the first two rules the transition regime occurs throughout a large time interval, and the information contained in the time series is transmitted, to a great extent, to it, as well as to the scaling regime. By using the third rule, on the contrary, the same information is essentially conveyed to the scaling regime, which, in fact, emerges very quickly after a fast transition process. We show that the DE method not only causes to emerge the long-range correlation with a given mu < 3, and so a basin of attraction different from the ordinary Gaussian one, but it also reveals the presence of memory effects induced by the time dependence of the solar flare rate. When this memory is annihilated by shuffling, the scaling parameter delta is shown to fit the theoretically expected function of mu. All this leads us to the compelling conclusion that mu = 2.138+/-0.01.

Journal Article↗

PubMed2003

Power-law distribution of pressure fluctuations in multiphase flow.

Bubbling fluidized beds are granular systems, in which a deep layer of particles is set in motion by a vertical gas stream, with the excess gas rising as bubbles through the bed. We show that pressure fluctuations in such a system have non-Gaussian statistics. The probability density function has a power-law drop-off and is very well represented by a Tsallis distribution. Its shape is explained through the folding of the Gaussian distribution of pressure fluctuations produced by a monodisperse set of bubbles, onto the actual distribution of bubble sizes in the bed, assuming that bubbles coalesce via a Smoluchowski-type aggregation process. Therefore, the Tsallis statistics arise as a result of bubble polydispersity, rather than system nonextensivity.

Journal Article↗

PubMed2001

The Gaussian derivative model for spatial-temporal vision: I. Cortical model.

How do we see the motion of objects as well as their shapes? The Gaussian Derivative (GD) spatial model is extended to time to help answer this question. The GD spatio-temporal model requires only two numbers to describe the complete three-dimensional space-time shapes of individual receptive fields in primate visual cortex. These two numbers are the derivative numbers along the respective spatial and temporal principal axes of a given receptive field. Nine transformation parameters allow for a standard geometric association of these intrinsic axes with the extrinsic environment. The GD spatio-temporal model describes in one framework the following properties of primate simple cell fields: motion properties, number of lobes in space-time, spatial orientation. location, and size. A discrete difference-of-offset-Gaussians (DOOG) model provides a plausible physiological mechanism to form GD-like model fields in both space and time. The GD model hypothesizes that receptive fields at the first stage of processing in the visual cortex approximate 'derivative analyzers' that estimate local spatial and temporal derivatives of the intensity profile in the visual environment. The receptive fields as modeled provide operators that can allow later stages of processing in either a biological or machine vision system to estimate the motion as well as the shapes of objects in the environment.

Animals↗

PubMed2004

Methods of bronchial tree reconstruction and camera distortion corrections for virtual endoscopic environments.

The use of three-dimensional visualization of anatomical structures in diagnostics and medical training is growing. The main components of virtual respiratory tract environments include reconstruction and simulation algorithms as well as correction methods of endoscope camera distortions in the case of virtually-enhanced navigation systems. Reconstruction methods rely usually on initial computer tomography (CT) image segmentation to trace contours of the tracheobronchial tree, which in turn are used in the visualization process. The main segmentation methods, including relatively simple approaches such as adaptive region-growing algorithms and more complex methods, e.g. hybrid algorithms based on region growing and mathematical morphology methods, are described in this paper. The errors and difficulties in the process of tracheobronchial tree reconstruction depend on the occurrence of distortions during CT image acquisition. They are usually related to the inability to exactly fulfil the sampling theorem's conditions. Other forms of distortions and noise such as additive white Gaussian noise, may also appear. The impact of these distortions on the segmentation and reconstruction may be diminished through the application of appropriately selected image prefiltering, which is also demonstrated in this paper. Methods of surface rendering (ray-casting, ray-tracing techniques) and volume rendering will be shown, with special focus on aspects of hardware and software implementations. Finally, methods of camera distortions correction and simulation are presented. The mathematical camera models, the scope of their applications and types of distortions were have also been indicated.

Algorithms↗

PubMed2005

Mechanism of HCS + O2 reaction: hydrogen- or oxygen-transfer?

In spite of the potential importance of the HCS radical in both combustion and interstellar processes, its chemical reactivity has not been tackled previously. In the present paper, the oxidation reaction of the HCS radical is theoretically investigated for the first time at the CCSD(T)/6-311++G(3df,2p)//BH&HLYP/6-311++G(d,p)+ZPVE and Gaussian-3//B3LYP/6-31G(d) levels. It is shown that the most feasible pathway is the O2 addition to the HCS radical forming the intermediate SC(H)OO which can undergo a subsequent O-extrusion leading to SC(H)O + 3O. This features an indirect O-transfer mechanism with the overall barrier of 4.4 and 3.5 kcal mol(-1), respectively, at the two levels. However, formation of the H-transfer product CS + HO2 is kinetically much less feasible, i.e., the direct mechanism has barriers of 14.3 and 8.7 kcal mol(-1), whereas the indirect mechanism has barriers of 12.6 and 10.7 kcal mol(-1), respectively. This result is in sharp contrast to the analogous HCO + O2 reaction, where the direct (with a barrier of 2.98 kcal mol(-1)) and indirect (2.26 kcal mol(-1)) H-transfer processes are highly competitive over the indirect O-transfer process (the least endothermicity is 19.9 kcal mol(-1)). The possible explanations and implications of the present results are provided.

Algorithms↗