并行子带HMM最大后验概率自适应非线性类估计算法

目前,自动语音识别(ASR)系统在实验室环境下获得了较高的识别率,但是在实际环境中,由于受到背景噪声和传输信道的影响,系统的识别性能急剧恶化.本文以听觉试验为基础,提出一种新的独立子带并行最大后验概率的非线性类估计算法,用以提高识别系统的鲁棒性.本算法利用多种噪声和识别内容功率谱差异,以及噪声在不同频带上对HMM影响的不同,采用多层感知机(MLP)对噪声环境下最大后验概率进行非线性映射,以减少识别系统由于环境不匹配而导致的识别性能下降.实验表明:该算法性能明显优于最大后验线性回归算法和Sangita提出的子带语音识别算法.     

A maximum a posteriori probability expectation maximization algorithm for image reconstruction in emission tomography

The expectation maximization method for maximum likelihood image reconstruction in emission tomography, based on the Poisson distribution of the statistically independent components of the image and measurement vectors, is extended to a maximum aposteriori image reconstruction using a multivariate Gaussian a priori probability distribution of the image vector. The approach is equivalent to a penalized maximum likelihood estimation with a special choice of the penalty function. The expectation maximization method is applied to find the a posteriori probability maximizer. A simple iterative formula is derived for a penalty function that is a weighted sum of the squared deviations of image vector components from their a priori mean values. The method is demonstrated to be superior to pure likelihood maximization, in that the penalty function prevents the occurrence of irregular high amplitude patterns in the image with a large number of iterations (the so-called "checkerboard effect" or "noise artifact").     

Precipitation type determination from spaceborne radar observations

The Tropical Rainfall Measuring Mission (TRMM) program is dedicated to observing and understanding the impact of tropical rainfall. Two important products of the TRMM precipitation radar include the classification of precipitation into convective and stratiform type as well as determination of the height of the brightband. Currently, TRMM uses an algorithm to arrive at these products based on a characterization of horizontal and vertical variability of reflectivity. This paper presents a new algorithm for classifying precipitation type into convective or stratiform as well as detect the height of brightband developed using wavelet analysis. The results obtained from wavelet analysis are compared against simultaneous ground validation products from the TRMM program. The results show that the wavelet-based algorithms provide fairly accurate results for both convective/stratiform classification as well as brightband determination.     

Maximum a posteriori spatial probability segmentation

An image segmentation algorithm that performs pixel-by-pixel segmentation on an image with consideration of the spatial information is described. The spatial information is the joint grey level values of the pixel to be segmented and its neighbouring pixels. The conditional probability that a pixel belongs to a particular class under the condition that the spatial information has been observed is defined to be the a posteriori spatial probability. A maximum a posteriori spatial probability (MASP) segmentation algorithm is proposed to segment an image such that each pixel is segmented into a particular class when the a posteriori spatial probability is a maximum. The proposed segmentation algorithm is implemented in an iterative form. During the iteration, a series of intermediate segmented images are produced among which the one that possesses the maximum amount of information in its spatial structure is chosen as the optimum segmented image. Results from segmenting synthetic and practical images demonstrate that the MASP algorithm is capable of achieving better results when compared with other global thresholding methods     

Modified maximum a posteriori decoding algorithm

A modified maximum a posteriori (MMAP) decoding algorithm that uses two extrinsic information values as the thresholds to determine the log-likelihood ratio, forward recursion probability and backward recursion probability is proposed. The MMAP requires less decoding time and complexity than an MAP decoder. Each probability of the proposed algorithm is derived and compared to that obtained using the MAP algorithm     

微波辐射计反演海表面温度和风场研究进展

海表面温度（SST）和海面风场是重要的海面气象水文参数,与海面的大部分物理过程关系密切。卫星遥感探测技术可以同时获得全球范围内的海表面温度和海面风场数据。其中,微波辐射计遥感海表面温度和风场具有不受云层遮挡的优势,本文就其发展过程作简要介绍。     

3-D maximum a posteriori estimation for single photon emission computed tomography on massively-parallel computers

A fully three-dimensional (3-D) implementation of the maximum a posteriori (MAP) method for single photon emission computed tomography (SPECT) is demonstrated. The 3-D reconstruction exhibits a major increase in resolution when compared to the generation of the series of separate 2-D slice reconstructions. As has been noted, the iterative EM algorithm for 2-D reconstruction is highly computational; the 3-D algorithm is far worse. To accommodate the computational complexity, previous work in the 2-D arena is extended, and an implementation on the class of massively parallel processors of the 3-D algorithm is demonstrated. Using a 16000- (4000-) processor MasPar/DECmpp-Sx machine, the algorithm is demonstrated to execute at 2.5 (7.8) s/EM-iteration for the entire 64x64x64 cube of 96 planar measurements obtained from the Siemens Orbiter rotating camera operating in the high-resolution mode.     

Simultaneous maximum a posteriori reconstruction of attenuation andactivity distributions from emission sinograms

In order to perform attenuation correction in emission tomography an attenuation map is required. We propose a new method to compute this map directly from the emission sinogram, eliminating the transmission scan from the acquisition protocol. The problem is formulated as an optimization task where the objective function is a combination of the likelihood and an a priori probability. The latter uses a Gibbs prior distribution to encourage local smoothness and a multimodal distribution for the attenuation coefficients. Since the attenuation process is different in positron emission tomography (PET) and single photon emission tomography (SPECT), a separate algorithm for each case is derived. The method has been tested on mathematical phantoms and on a few clinical studies. For PET, good agreement was found between the images obtained with transmission measurements and those produced by the new algorithm in an abdominal study. For SPECT, promising simulation results have been obtained for nonhomogeneous attenuation due to the presence of the lungs.     

Maximum a posteriori estimation of multichannel Bernoulli-Gaussiansequences

General problems and solutions are described for maximum a posteriori estimation of multichannel Bernoulli-Gaussian sequences, which are inputs to a linear discrete-time multivariable system. The authors first develop a separation principle, which indicates that one can estimate multichannel Gaussian amplitudes and Bernoulli events separately. They then discuss approaches for estimation of these quantities     

Parallel algorithms for maximum a posteriori estimation of spin density and spin-spin decay in magnetic resonance imaging

A maximum a posteriori (MAP) algorithm is presented for the estimation of spin-density and spin-spin decay distributions from frequency and phase-encoded magnetic resonance imaging data. Linear spatial localization gradients are assumed: the y-encode gradient applied during the phase preparation time of duration tau before measurement collection, and the x-encode gradient applied during the full data collection time t>/=0. The MRI signal model developed in M.I. Miller et al., J. Magn. Reson., ser. B (Apr. 1995) is used in which a signal resulting from M phase encodes (rows) and N frequency encode dimensions (columns) is modeled as a superposition of MN sinc-modulated exponentially decaying sinusoids with unknown spin-density and spin-spin decay parameters. The nonlinear least-squares MAP estimate of the spin density and spin-spin decay distributions solves for the 2MN spin-density and decay parameters minimizing the squared-error between the measured data and the sine-modulated exponentially decay signal model using an iterative expectation-maximization algorithm. A covariance diagonalizing transformation is derived which decouples the joint estimation of MN sinusoids into M separate N sinusoid optimizations, yielding an order of magnitude speed up in convergence. The MAP solutions are demonstrated to deliver a decrease in standard deviation of image parameter estimates on brain phantom data of greater than a factor of two over Fourier-based estimators of the spin density and spin-spin decay distributions. A parallel processor implementation is demonstrated which maps the N sinusoid coupled minimization to separate individual simple minimizations, one for each processor.     

Solving the equation for the maximum a posteriori estimate

An iterative technique is given for solving the equation for the maximum a posteriori estimate of the modulation on a carrier which is combined with noise. The method has been tested in the FM case with satisfactory results.     

Maximum-likelihood versus maximum a posteriori parameter estimation of physiological system models: the C-peptide impulse response case study

Maximum-likelihood (ML), also given its connection to least-squares (LS), is widely adopted in parameter estimation of physiological system models, i.e., assigning numerical values to the unknown model parameters from the experimental data. A more sophisticated but less used approach is maximum a posteriori (MAP) estimation. Conceptually, while ML adopts a Fisherian approach, i.e., only experimental measurements are supplied to the estimator, MAP estimation is a Bayesian approach, i.e., a priori available statistical information on the unknown parameters is also exploited for their estimation. In this paper, after a brief review of the theory behind ML and MAP estimators, we compare their performance in the solution of a case study concerning the determination of the parameters of a sum of exponential model which describes the impulse response of C-peptide (CP), a key substance for reconstructing insulin secretion. The results show that MAP estimation always leads to parameter estimates with a precision (sometimes significantly) higher than that obtained through ML, at the cost of only a slightly worse fit. Thus, a three exponential model can be adopted to describe the CP impulse response model in place of the two exponential model usually identified in the literature by the ML/LS approach. Simulated case studies are also reported to evidence the importance of taking into account a priori information in a data poor situation, e.g., when a few or too noisy measurements are available. In conclusion, our results show that, when a priori information on the unknown model parameters is available, Bayes estimation can be of relevant interest, since it can significantly improve the precision of parameter estimates with respect to Fisher estimation. This may also allow the adoption of more complex models than those determinable by a Fisherian approach.     

Generalized maximum a posteriori processing of multichannel images and applications

This paper considers several aspects of robust estimation in the restoration of mutichannel images. Robust functionals emerging from a generalized maximum a posteriori (MAP) approach are employed for the representation of both the noise and the signal statistics. Several linear multichannel techniques can be derived as special cases of the approach presented. In addition, the robust approach derives nonlinear algorithms that simultaneously account for the suppression of nominal noise and outliers, and for the efficient reconstruction of sharp detailed structure in the estimate. The robust multichannel approach is presented as a general approach for the regularization of the ill-posed restoration problem. From this perspective, we develop a method for the selection of the regularization parameter, which can be used in a wide variety of applications that may or may not involve noise outliers. We consider several issues associated with the application of robust algorithms to multichannel images, we discuss computational inefficiencies of such algorithms, and we propose approximations that are appropriate for their cost-efficient multichannel implementation. We demonstrate the robust approach in two examples from the rapidly developing fields of color image processing and multiresolution image processing in the wavelet domain.The research presented in this paper was partially supported by the Graduate School of the University of Minnesota under Summer Research Fellowship no. 15556 and Grant-in-Aid Award no. 15987.     

Compensation of elevation angle variations in polarimetricbrightness temperature measurements from airborne microwave radiometers

This paper presents a method for compensating the elevation angle fluctuations occurring in airborne radiometry due to aircraft roll and pitch. The correction is based on a radiative transfer model, and is demonstrated by real data from conical scans over the ocean, showing good results     

Simulated annealing for maximum a posteriori parameter estimationof hidden Markov models

Hidden Markov models are mixture models in which the populations from one observation to the next are selected according to an unobserved finite state-space Markov chain. Given a realization of the observation process, our aim is to estimate both the parameters of the Markov chain and of the mixture model in a Bayesian framework. We present an original simulated annealing algorithm which, in the same way as the EM (expectation-maximization) algorithm, relies on data augmentation, and is based on stochastic simulation of the hidden Markov chain. This algorithm is shown to converge toward the set of maximum a posteriori (MAP) parameters under suitable regularity conditions     

Super-resolution based on fast registration and maximum a posteriori reconstruction.

In this paper, we propose a maximum a posteriori ramework for the super-resolution problem, i.e., reconstructing high-resolution images from shifted, rotated, low-resolution degraded observations. The main contributions of this work are two; first, the use of a new locally adaptive edge preserving prior for the super-resolution problem. Second an efficient two-step reconstruction methodology that includes first an initial registration using only the low-resolution degraded observations. This is followed by a fast iterative algorithm implemented in the discrete Fourier transform domain in which the restoration, interpolation and the registration subtasks of this problem are preformed simultaneously. We present examples with both synthetic and real data that demonstrate the advantages of the proposed framework.     

Convergence of the maximum a posteriori path estimator in hiddenMarkov models

In a hidden Markov model (HMM) the underlying finite-state Markov chain cannot be observed directly but only by an additional process. We are interested in estimating the unknown path of the Markov chain. The most widely used estimator is the maximum a posteriori path estimator (MAP path estimator). It can be calculated effectively by the Viterbi (1967) algorithm as is, e.g., frequently done in the field of coding theory, correction of intersymbol interference, and speech recognition. We investigate (component-wise) convergence of the MAP path estimator. Convergence is shown under the condition of unbounded likelihood ratios. This condition is satisfied in the important case of HMMs with additive white Gaussian noise. We also prove convergence, if the Markov chain has two states. The so-called Viterbi paths are an important tool for obtaining these results     

Retrieval of atmospheric and surface parameters from satellite microwave radiometers over the Mediterranean Sea

A procedure to estimate atmospheric and sea surface parameters in the Mediterranean area from satellite microwave radiometric measurements is described. The method is founded on a simulator of brightness temperatures at the top of the atmosphere. The simulator is based on microwave sea emissivity and scattering model functions, derived from the outputs of the SEAWIND software, which implements a two-scale microwave sea surface model and a radiative transfer scheme in a nonscattering atmosphere. The development of the model functions aims to reduce the SEAWIND computational time, still maintaining its sensitivity to the main geophysical variables. Different adaptations of the simulation model have been performed to better reproduce the radiometric data in the region of interest. A comparison between the simulations and the Special Sensor Microwave/Imager (SSM/I) observations acquired throughout year 2000 over the Mediterranean Sea has permitted us to refine the model functions as well as to assess the whole simulation procedure. As for the inversion problem, a regression analysis has been applied to two different synthetic datasets to retrieve integrated precipitable water vapor, liquid water path and wind speed. The first dataset simulates the observations of SSM/I, whilst the second one concerns the Advanced Microwave Scanning Radiometer for the Earth Observing System (AMSR-E). Both have been generated by using the ECMWF atmospheric profiles and the measurements of the SeaWinds scatterometer aboard QuickSCAT. The SSM/I data have been used to carry out a statistical validation of the estimators. AMSR-E observations of a Tramontane-Mistral event, typical of the Mediterranean Sea, have been analyzed to evaluate the benefits of its expanded channel capability.     

Large-signal microwave characterization of AlGaAs/GaAs HBT's basedon a physics-based electrothermal model

A physics-based multicell electrothermal equivalent circuit model is described that is applied to the large-signal microwave characterization of AlGaAs/GaAs HBT's. This highly efficient model, which incorporates a new multifinger electrothermal model, has been used to perform dc, small-signal and load-pull characterization, and investigate parameter-spreads due to fabrication process variations. An enhanced Newton algorithm is presented for solving the nonlinear system of equations for the model and associated circuit simulator, which allows a faster and more robust solution than contemporary quasi-Newton nonlinear schemes. The model has been applied to the characterization of heterojunction bipolar transistor (HBT) microwave power amplifiers