Showing papers on "Poisson distribution published in 2009"

Interference in Large Wireless Networks

Abstract: Since interference is the main performance-limiting factor in most wireless networks, it is crucial to characterize the interference statistics. The two main determinants of the interference are the network geometry (spatial distribution of concurrently transmitting nodes) and the path loss law (signal attenuation with distance). For certain classes of node distributions, most notably Poisson point processes, and attenuation laws, closed-form results are available, for both the interference itself as well as the signal-to-interference ratios, which determine the network performance. This monograph presents an overview of these results and gives an introduction to the analytical techniques used in their derivation. The node distribution models range from lattices to homogeneous and clustered Poisson models to general motion-invariant ones. The analysis of the more general models requires the use of Palm theory, in particular conditional probability generating functionals, which are briefly introduced in the appendix.

The Analysis of Count Data : A Gentle Introduction to Poisson Regression and Its Alternatives

Abstract: Count data reflect the number of occurrences of a behavior in a fixed period of time (e.g., number of aggressive acts by children during a playground period). In cases in which the outcome variable is a count with a low arithmetic mean (typically < 10), standard ordinary least squares regression may produce biased results. We provide an introduction to regression models that provide appropriate analyses for count data. We introduce standard Poisson regression with an example and discuss its interpretation. Two variants of Poisson regression, overdispersed Poisson regression and negative binomial regression, are introduced that may provide more optimal results when a key assumption of standard Poisson regression is violated. We also discuss the problems of excess zeros in which a subgroup of respondents who would never display the behavior are included in the sample and truncated zeros in which respondents who have a zero count are excluded by the sampling plan. We provide computer syntax for our illustrat...

On the Specification of the Gravity Model of Trade: Zeros, Excess Zeros and Zero-inflated Estimation

Abstract: Conventional studies of bilateral trade patterns specify a log-normal gravity equation for empirical estimation. However, the log-normal gravity equation suffers from three problems: the bias created by the logarithmic transformation, the failure of the homoscedasticity assumption, and the way zero values are treated. These problems normally result in biased and inefficient estimates. Recently, the Poisson specification of the trade gravity model has received attention as an alternative to the log-normality assumption (Santos Silva and Tenreyro, 2006). However, the standard Poisson model is vulnerable for problems of overdispersion and excess zero flows. To overcome these problems, this paper considers modified Poisson fixed-effects estimations (negative binomial, zero-inflated). Extending the empirical model put forward by Santos Silva and Tenreyro (2006), we show how these techniques may provide viable alternatives to both the log-normal and standard Poisson specification of the gravity model of trade.

Bayesian Poisson regression for crowd counting

Abstract: Poisson regression models the noisy output of a counting function as a Poisson random variable, with a log-mean parameter that is a linear function of the input vector In this work, we analyze Poisson regression in a Bayesian setting, by introducing a prior distribution on the weights of the linear function Since exact inference is analytically unobtainable, we derive a closed-form approximation to the predictive distribution of the model We show that the predictive distribution can be kernelized, enabling the representation of non-linear log-mean functions We also derive an approximate marginal likelihood that can be optimized to learn the hyperparameters of the kernel We then relate the proposed approximate Bayesian Poisson regression to Gaussian processes Finally, we present experimental results using Bayesian Poisson regression for crowd counting from low-level features

Interference and Outage in Clustered Wireless Ad Hoc Networks

Abstract: In the analysis of large random wireless networks, the underlying node distribution is almost ubiquitously assumed to be the homogeneous Poisson point process. In this paper, the node locations are assumed to form a Poisson cluster process on the plane. We derive the distributional properties of the interference and provide upper and lower bounds for its distribution. We consider the probability of successful transmission in an interference-limited channel when fading is modeled as Rayleigh. We provide a numerically integrable expression for the outage probability and closed-form upper and lower bounds. We show that when the transmitter-receiver distance is large, the success probability is greater than that of a Poisson arrangement. These results characterize the performance of the system under geographical or MAC-induced clustering. We obtain the maximum intensity of transmitting nodes for a given outage constraint, i.e., the transmission capacity (of this spatial arrangement) and show that it is equal to that of a Poisson arrangement of nodes. For the analysis, techniques from stochastic geometry are used, in particular the probability generating functional of Poisson cluster processes, the Palm characterization of Poisson cluster processes, and the Campbell-Mecke theorem.

Image deblurring with Poisson data: from cells to galaxies

Abstract: Image deblurring is an important topic in imaging science. In this review, we consider together fluorescence microscopy and optical/infrared astronomy because of two common features: in both cases the imaging system can be described, with a sufficiently good approximation, by a convolution operator, whose kernel is the so-called point-spread function (PSF); moreover, the data are affected by photon noise, described by a Poisson process. This statistical property of the noise, that is common also to emission tomography, is the basis of maximum likelihood and Bayesian approaches introduced in the mid eighties. From then on, a huge amount of literature has been produced on these topics. This review is a tutorial and a review of a relevant part of this literature, including some of our previous contributions. We discuss the mathematical modeling of the process of image formation and detection, and we introduce the so-called Bayesian paradigm that provides the basis of the statistical treatment of the problem. Next, we describe and discuss the most frequently used algorithms as well as other approaches based on a different description of the Poisson noise. We conclude with a review of other topics related to image deblurring such as boundary effect correction, space-variant PSFs, super-resolution, blind deconvolution and multiple-image deconvolution.

Density Estimation by Spatially Explicit Capture–Recapture: Likelihood-Based Methods

Abstract: Population density is a key ecological variable, and it has recently been shown how captures on an array of traps over several closely-spaced time intervals may be modelled to provide estimates of population density (Borchers and Efford 2007). Specifics of the model depend on the properties of the traps (more generally ‘detectors’). We provide a concise description of the newly developed likelihood-based methods and extend them to include ‘proximity detectors’ that do not restrict the movements of animals after detection. This class of detector includes passive DNA sampling and camera traps. The probability model for spatial detection histories comprises a submodel for the distribution of home-range centres (e.g. 2-D Poisson) and a detection submodel (e.g. halfnormal function of distance between a range centre and a trap). The model may be fitted by maximising either the full likelihood or the likelihood conditional on the number of animals observed. A wide variety of other effects on detection probability may be included in the likelihood using covariates or mixture models, and differences in density between sites or between times may also be modelled. We apply the method to data on stoats Mustela erminea in a New Zealand beech forest identified by microsatellite DNA from hair samples. The method assumes that multiple individuals may be recorded at a detector on one occasion. Formal extension to ‘single-catch’ traps is difficult, but in our simulations the ‘multi-catch’ model yielded nearly unbiased estimates of density for moderate levels of trap saturation (≤ 86% traps occupied), even when animals were clustered or the traps spanned a gradient in density.

Fractional Poisson processes and related planar random motions

Abstract: We present three different fractional versions of the Poisson process and some related results concerning the distribution of order statistics and the compound Poisson process. The main version is constructed by considering the difference-differential equation governing the distribution of the standard Poisson process, $ N(t),t>0$, and by replacing the time-derivative with the fractional Dzerbayshan-Caputo derivative of order $ u\in(0,1]$. For this process, denoted by $\mathcal{N}_ u(t),t>0,$ we obtain an interesting probabilistic representation in terms of a composition of the standard Poisson process with a random time, of the form $\mathcal{N}_ u(t)= N(\mathcal{T}_{2 u}(t)),$ $t>0$. The time argument $\mathcal{T}_{2 u }(t),t>0$, is itself a random process whose distribution is related to the fractional diffusion equation. We also construct a planar random motion described by a particle moving at finite velocity and changing direction at times spaced by the fractional Poisson process $\mathcal{N}_ u.$ For this model we obtain the distributions of the random vector representing the position at time $t$, under the condition of a fixed number of events and in the unconditional case. For some specific values of $ u\in(0,1]$ we show that the random position has a Brownian behavior (for $ u =1/2$) or a cylindrical-wave structure (for $ u =1$).

A Proximal Iteration for Deconvolving Poisson Noisy Images Using Sparse Representations

Abstract: We propose an image deconvolution algorithm when the data is contaminated by Poisson noise. The image to restore is assumed to be sparsely represented in a dictionary of waveforms such as the wavelet or curvelet transforms. Our key contributions are as follows. First, we handle the Poisson noise properly by using the Anscombe variance stabilizing transform leading to a nonlinear degradation equation with additive Gaussian noise. Second, the deconvolution problem is formulated as the minimization of a convex functional with a data-fidelity term reflecting the noise properties, and a nonsmooth sparsity-promoting penalty over the image representation coefficients (e.g., lscr1 -norm). An additional term is also included in the functional to ensure positivity of the restored image. Third, a fast iterative forward-backward splitting algorithm is proposed to solve the minimization problem. We derive existence and uniqueness conditions of the solution, and establish convergence of the iterative algorithm. Finally, a GCV-based model selection procedure is proposed to objectively select the regularization parameter. Experimental results are carried out to show the striking benefits gained from taking into account the Poisson statistics of the noise. These results also suggest that using sparse-domain regularization may be tractable in many deconvolution applications with Poisson noise such as astronomy and microscopy.

Further Simulation Evidence on the Performance of the Poisson Pseudo-Maximum Likelihood Estimator

Abstract: We extend the simulation results given in Santos-Silva and Tenreyro (2006, 'The Log of Gravity', The Review of Economics and Statistics, 88, pp.641-658) by considering data generated as a finite mixture of gamma variates. Data generated in this way can naturally have a large proportion of zeros and is fully compatible with constant elasticity models such as the gravity equation. Our results confirm that the Poisson pseudo maximum likelihood estimator is generally well behaved.

Two-step estimation for inhomogeneous spatial point processes

Abstract: Summary The paper is concerned with parameter estimation for inhomogeneous spatial point processes with a regression model for the intensity function and tractable second-order properties (K-function) Regression parameters are estimated by using a Poisson likelihood score estimating function and in the second step minimum contrast estimation is applied for the residual clustering parameters Asymptotic normality of parameter estimates is established under certain mixing conditions and we exemplify how the results may be applied in ecological studies of rainforests

Bayesian Multivariate Poisson Lognormal Models for Crash Severity Modeling and Site Ranking

Abstract: Traditionally, highway safety analyses have used univariate Poisson or negative binomial distributions to model crash counts for different levels of crash severity. Because unobservables or omitted variables are shared across severity levels, however, crash counts are multivariate in nature. This research uses full Bayes multivariate Poisson lognormal models to estimate the expected crash frequency for different levels of crash severity and then compares those estimates to independent or univariate Poisson lognormal estimates. The multivariate Poisson lognormal model fits better than the univariate model and improves the precision in crash-frequency estimates. The covariances and correlations among crash severities are high (correlations range from 0.47 to 0.97), with the highest values found between contiguous severity levels. Considering this correlation between severity levels improves the precision of the expected number of crashes. The multivariate estimates are used with cost data from the Pennsylvania Department of Transportation to develop the expected crash cost (and excess expected cost) per segment, which is then used to rank sites for safety improvements. The multivariate-based top-ranked segments are found to have consistently higher costs and excess costs than the univariate estimates, which is due to higher multivariate estimates of fatalities and major injuries (due to the random effects parameter). These higher estimated frequencies, in turn, produce different rankings for the multivariate and independent models. The finding of a high correlation between contiguous severity levels is consistent with some of the literature, but additional tests of multivariate models are recommended. The improved precision has important implications for the identification of sites with promise (SWiPs), because one formulation includes the standard deviation of crash frequencies for similar sites as part of the assessment of SWiPs.

Modeling abundance using N-mixture models: the importance of considering ecological mechanisms

Abstract: Predicting abundance across a species' distribution is useful for studies of ecology and biodiversity management. Modeling of survey data in relation to environmental variables can be a powerful method for extrapolating abundances across a species' distribution and, consequently, calculating total abundances and ultimately trends. Research in this area has demonstrated that models of abundance are often unstable and produce spurious estimates, and until recently our ability to remove detection error limited the development of accurate models. The N-mixture model accounts for detection and abundance simultaneously and has been a significant advance in abundance modeling. Case studies that have tested these new models have demonstrated success for some species, but doubt remains over the appropriateness of standard N-mixture models for many species. Here we develop the N-mixture model to accommodate zero-inflated data, a common occurrence in ecology, by employing zero-inflated count models. To our knowledge, this is the first application of this method to modeling count data. We use four variants of the N-mixture model (Poisson, zero-inflated Poisson, negative binomial, and zero-inflated negative binomial) to model abundance, occupancy (zero-inflated models only) and detection probability of six birds in South Australia. We assess models by their statistical fit and the ecological realism of the parameter estimates. Specifically, we assess the statistical fit with AIC and assess the ecological realism by comparing the parameter estimates with expected values derived from literature, ecological theory, and expert opinion. We demonstrate that, despite being frequently ranked the “best model” according to AIC, the negative binomial variants of the N-mixture often produce ecologically unrealistic parameter estimates. The zero-inflated Poisson variant is preferable to the negative binomial variants of the N-mixture, as it models an ecological mechanism rather than a statistical phenomenon and generates reasonable parameter estimates. Our results emphasize the need to include ecological reasoning when choosing appropriate models and highlight the dangers of modeling statistical properties of the data. We demonstrate that, to obtain ecologically realistic estimates of abundance, occupancy and detection probability, it is essential to understand the sources of variation in the data and then use this information to choose appropriate error distributions. Copyright ESA. All rights reserved.

On the Capacity of the Discrete-Time Poisson Channel

Abstract: The large-inputs asymptotic capacity of a peak-power and average-power limited discrete-time Poisson channel is derived using a new firm (nonasymptotic) lower bound and an asymptotic upper bound. The upper bound is based on the dual expression for channel capacity and the notion of capacity-achieving input distributions that escape to infinity. The lower bound is based on a lower bound on the entropy of a conditionally Poisson random variable in terms of the differential entropy of its conditional mean.

Non-life insurance mathematics : an introduction with the Poisson process

Abstract: Collective Risk Models.- The Basic Model.- Models for the Claim Number Process.- The Total Claim Amount.- Ruin Theory.- Experience Rating.- Bayes Estimation.- Linear Bayes Estimation.- A Point Process Approach to Collective Risk Theory.- The General Poisson Process.- Poisson Random Measures in Collective Risk Theory.- Weak Convergence of Point Processes.- Special Topics.- An Excursion to L#x00E9 vy Processes.- Cluster Point Processes.

Regularization parameter selection methods for ill-posed Poisson maximum likelihood estimation

Abstract: In image processing applications, image intensity is often measured via the counting of incident photons emitted by the object of interest. In such cases, image data noise is accurately modeled by a Poisson distribution. This motivates the use of Poisson maximum likelihood estimation for image reconstruction. However, when the underlying model equation is ill-posed, regularization is needed. Regularized Poisson likelihood estimation has been studied extensively by the authors, though a problem of high importance remains: the choice of the regularization parameter. We will present three statistically motivated methods for choosing the regularization parameter, and numerical examples will be presented to illustrate their effectiveness.

Estimating Abundance Using Mark–Resight When Sampling Is with Replacement or the Number of Marked Individuals Is Unknown

Abstract: Although mark-resight methods can often be a less expensive and less invasive means for estimating abundance in long-term population monitoring programs, two major limitations of the estimators are that they typically require sampling without replacement and/or the number of marked individuals available for resighting to be known exactly. These requirements can often be difficult to achieve. Here we address these limitations by introducing the Poisson log and zero-truncated Poisson log-normal mixed effects models (PNE and ZPNE, respectively). The generalized framework of the models allow the efficient use of covariates in modeling resighting rate and individual heterogeneity parameters, information-theoretic model selection and multimodel inference, and the incorporation of individually unidentified marks. Both models may be implemented using standard statistical computing software, but they have also been added to the mark-recapture freeware package Program MARK. We demonstrate the use and advantages of (Z)PNE using black-tailed prairie dog data recently collected in Colorado. We also investigate the expected relative performance of the models in simulation experiments. Compared to other available estimators, we generally found (Z)PNE to be more precise with little or no loss in confidence interval coverage. With the recent introduction of the logit-normal mixed effects model and (Z)PNE, a more flexible and efficient framework for mark-resight abundance estimation is now available for the sampling conditions most commonly encountered in these studies.

The fracture toughness of composite laminates with a negative Poisson's ratio

Abstract: Two sets of carbon fibre laminate configurations have been designed, manufactured and examined. The first consisted of one in-plane auxetic (i.e. having a negative Poisson's ratio) and one with matched elastic moduli, but a positive in-plane Poisson's ratio. The second consisted of three laminates, again with matched elastic moduli, but with negative, near zero and positive thru-thickness Poisson's ratio. Fracture toughness is ; predicted to be enhanced if a negative thru-thickness Poisson's ratio is present. This study also shows that more energy is required to propagate a crack in the auxetic laminate and that it is less notch sensitive than its conventional counterpart for the negative in-plane Poisson's ratio.

Feature selection with a measure of deviations from Poisson in text categorization

Abstract: To improve the performance of automatic text classification, it is desirable to reduce a high dimensionality of the feature space. In this paper, we propose a new measure for selecting features, which estimates term importance based on how largely the probability distribution of each term deviates from the standard Poisson distribution. In information retrieval literatures, the deviation from Poisson has been used as a measure for weighting keywords and this motivates us to adopt the deviation from Poisson as a measure for feature selection in text classification tasks. The proposed measure is constructed so as to have the same computational complexity with other standard measures used for feature selection. To test the effectiveness of our method, we conducted evaluation experiments on Reuters-21578 corpus with support vector machine and k-NN classifiers. In the experiments, we performed binary classifications to determine whether each of the test documents belongs to a certain target category or not. For the target category, each of the top 10 categories of Reuters-21578 was used because of enough numbers of training and test documents. Four measures were used for feature selection; information gain (IG), @g^2-statistic, Gini index and the proposed measure in this work. Both the proposed measure and Gini index proved to be better than IG and @g^2-statistic in terms of macro-averaged and micro-averaged values of F"1, especially at higher vocabulary reduction levels.

GLM and GAM for Count Data

Abstract: A generalised linear model (GLM) or a generalised additive model (GAM) consists of three steps: (i) the distribution of the response variable, (ii) the specification of the systematic component in terms of explanatory variables, and (iii) the link between the mean of the response variable and the systematic part. In Chapter 8, we discussed several different distributions for the response variable: Normal, Poisson, negative binomial, geometric, gamma, Bernoulli, and binomial distributions. One of these distributions can be used for the first step mentioned above. In fact, later in Chapter 11, we see how you can also use a mixture of two distributions for the response variable; but in this chapter, we only work with one distribution at a time.

Analyzing short-term noise dependencies of spike-counts in macaque prefrontal cortex using copulas and the flashlight transformation.

Abstract: Simultaneous spike-counts of neural populations are typically modeled by a Gaussian distribution. On short time scales, however, this distribution is too restrictive to describe and analyze multivariate distributions of discrete spike-counts. We present an alternative that is based on copulas and can account for arbitrary marginal distributions, including Poisson and negative binomial distributions as well as second and higher-order interactions. We describe maximum likelihood-based procedures for fitting copula-based models to spike-count data, and we derive a so-called flashlight transformation which makes it possible to move the tail dependence of an arbitrary copula into an arbitrary orthant of the multivariate probability distribution. Mixtures of copulas that combine different dependence structures and thereby model different driving processes simultaneously are also introduced. First, we apply copula-based models to populations of integrate-and-fire neurons receiving partially correlated input and show that the best fitting copulas provide information about the functional connectivity of coupled neurons which can be extracted using the flashlight transformation. We then apply the new method to data which were recorded from macaque prefrontal cortex using a multi-tetrode array. We find that copula-based distributions with negative binomial marginals provide an appropriate stochastic model for the multivariate spike-count distributions rather than the multivariate Poisson latent variables distribution and the often used multivariate normal distribution. The dependence structure of these distributions provides evidence for common inhibitory input to all recorded stimulus encoding neurons. Finally, we show that copula-based models can be successfully used to evaluate neural codes, e.g., to characterize stimulus-dependent spike-count distributions with information measures. This demonstrates that copula-based models are not only a versatile class of models for multivariate distributions of spike-counts, but that those models can be exploited to understand functional dependencies.

Statistical Methods in Practice: For Scientists and Technologists

Abstract: Preface. 1 Samples and populations. Introduction. What a lottery! No can do. Nobody is listening to me. How clean is my river? Discussion. 2 What is the true mean? Introduction. Presenting data. Averages. Measures of variability. Relative standard deviation . Degrees of freedom. Confidence interval for the population mean. Sample sizes. How much moisture is in the raw material? Problems. 3 Exploratory data analysis. Introduction. Histograms: is the process capable of meeting specifications? Box plots: how long before the lights go out? The box plot in practice. Problems. 4 Significance testing. Introduction. The one-sample t -test. The significance testing procedure. Confidence intervals as an alternative to significance testing. Confidence interval for the population standard deviation. F -test for ratio of standard deviations. Problems. 5 The normal distribution. Introduction. Properties of the normal distribution. Example. Setting the process mean. Checking for normality. Uses of the normal distribution. Problems. 6 Tolerance intervals. Introduction. Example. Confidence intervals and tolerance intervals. 7 Outliers. Introduction. Grubbs' test. Warning. 8 Significance tests for comparing two means. Introduction. Example: watching paint lose its gloss. The two-sample t -test for independent samples. An alternative approach: a confidence intervals for the difference between population means. Sample size to estimate the difference between two means. A production example. Confidence intervals for the difference between the two suppliers. Sample size to estimate the difference between two means. Conclusions. Problems. 9 Significance tests for comparing paired measurements. Introduction. Comparing two fabrics. The wrong way. The paired sample t -test. Presenting the results of significance tests. One-sided significance tests. Problems. 10 Regression and correlation. Introduction. Obtaining the best straight line. Confidence intervals for the regression statistics. Extrapolation of the regression line. Correlation coefficient. Is there a significant relationship between the variables? How good a fit is the line to the data? Assumptions. Problems. 11 The binomial distribution. Introduction. Example. An exact binomial test. A quality assurance example. What is the effect of the batch size? Problems. 12 The Poisson distribution. Introduction. Fitting a Poisson distribution. Are the defects random? The Poisson distribution. Poisson dispersion test. Confidence intervals for a Poisson count. A significance test for two Poisson counts. How many black specks are in the batch? How many pathogens are there in the batch? Problems. 13 The chi-squared test for contingency tables. Introduction. Two-sample test for percentages. Comparing several percentages. Where are the differences? Assumptions. Problems. 14 Non-parametric statistics. Introduction. Descriptive statistics. A test for two independent samples: Wilcoxon-Mann-Whitney test. A test for paired data: Wilcoxon matched-pairs sign test. What type of data can be used? Example: cracking shoes. Problems. 15 Analysis of variance: Components of variability. Introduction. Overall variability. Analysis of variance. A practical example. Terminology. Calculations. Significance test. Variation less than chance? When should the above methods not be used? Between- and within-batch variability. How many batches and how many prawns should be sampled? Problems. 16 Cusum analysis for detecting process changes. Introduction. Analysing past data. Intensity. Localised standard deviation. Significance test. Yield. Conclusions from the analysis. Problem. 17 Rounding of results. Introduction. Choosing the rounding scale. Reporting purposes: deciding the amount of rounding. Reporting purposes: rounding of means and standard deviations. Recording the original data and using means and standard deviations in statistical analysis. References. Solutions to Problems. Statistical Tables. Index .

Number of Accidents or Number of Claims? An Approach with Zero-Inflated Poisson Models for Panel Data

Abstract: P>The hunger for bonus is a well-known phenomenon in insurance, meaning that the insured does not report all of his accidents to save bonus on his next year's premium. In this article, we assume that the number of accidents is based on a Poisson distribution but that the number of claims is generated by censorship of this Poisson distribution. Then, we present new models for panel count data based on the zero-inflated Poisson distribution. From the claims distributions, we propose an approximation of the accident distribution, which can provide insight into the behavior of insureds. A numerical illustration based on the reported claims of a Spanish insurance company is included to support this discussion.

Estimation methods for the discrete Poisson–Lindley distribution

Abstract: In this paper, we show that the method of moments and maximum likelihood estimators of the parameter of the discrete Poisson–Lindley distribution are consistent and asymptotically normal. The asymptotic variances of the two estimators are almost equal, indicating that the two estimators are almost equally efficient. Also, a simulation study is presented to compare the two estimators. Finally, the two estimators, their standard errors, and the confidence intervals are compared for two published data sets.

Model uncertainty in claims reserving within Tweedie's compound Poisson models

Abstract: In this paper we examine the claims reserving problem using Tweedie's compound Poisson model. We develop the maximum likelihood and Bayesian Markov chain Monte Carlo simulation approaches to fit the model and then compare the estimated models under different scenarios. The key point we demonstrate relates to the comparison of reserving quantities with and without model uncertainty incorporated into the prediction. We consider both the model selection problem and the model averaging solutions for the predicted reserves. As a part of this process we also consider the sub problem of variable selection to obtain a parsimonious representation of the model being fitted.

Bayesian Inference on Multiscale Models for Poisson Intensity Estimation: Applications to Photon-Limited Image Denoising

Abstract: We present an improved statistical model for analyzing Poisson processes, with applications to photon-limited imaging. We build on previous work, adopting a multiscale representation of the Poisson process in which the ratios of the underlying Poisson intensities (rates) in adjacent scales are modeled as mixtures of conjugate parametric distributions. Our main contributions include: 1) a rigorous and robust regularized expectation-maximization (EM) algorithm for maximum-likelihood estimation of the rate-ratio density parameters directly from the noisy observed Poisson data (counts); 2) extension of the method to work under a multiscale hidden Markov tree model (HMT) which couples the mixture label assignments in consecutive scales, thus modeling interscale coefficient dependencies in the vicinity of image edges; 3) exploration of a 2-D recursive quad-tree image representation, involving Dirichlet-mixture rate-ratio densities, instead of the conventional separable binary-tree image representation involving beta-mixture rate-ratio densities; and 4) a novel multiscale image representation, which we term Poisson-Haar decomposition, that better models the image edge structure, thus yielding improved performance. Experimental results on standard images with artificially simulated Poisson noise and on real photon-limited images demonstrate the effectiveness of the proposed techniques.

Maximal regularity for stochastic convolutions driven by Lévy processes

Abstract: We generalize the maximal regularity result from Da Prato and Lunardi (Atti Accad Naz Lincei Cl Sci Fis Mat Natur Rend Lincei (9) Mat Appl 9(1):25–29, 1998) to stochastic convolutions driven by time homogenous Poisson random measures and cylindrical infinite dimensional Wiener processes.