第2版-图像分析.随机场和马尔可夫链蒙特卡罗方法 内容简介

该书主要研究了图像分析的随机场方法、相关的马尔可夫链蒙特卡罗法、贝叶斯图像分析的统计推断，重点关注了一般性原理，具体的应用细节稍微少点。该书可以说是为数学专业、统计学、物理学、工程和计算机专业的学生和专家量身定做的，书中更多的体现了一些数学方法，而非综述。基本上不需要读者拥有更深的数学或者统计学的知识。第2版改动很大，许多数据都做了更新，并且新增了似然函数的精确抽样和全局*优法。

第2版-图像分析.随机场和马尔可夫链蒙特卡罗方法 本书特色

该书主要研究了图像分析的随机场方法、相关的马尔可夫链蒙特卡罗法、贝叶斯图像分析的统计推断，重点关注了一般性原理，具体的应用细节稍微少点。该书可以说是为数学专业、统计学、物理学、工程和计算机专业的学生和专家量身定做的，书中更多的体现了一些数学方法，而非综述。基本上不需要读者拥有更深的数学或者统计学的知识。第2版改动很大，许多数据都做了更新，并且新增了似然函数的精确抽样和全局*优法。

第2版-图像分析.随机场和马尔可夫链蒙特卡罗方法 目录

IntroductionPart Ⅰ．Bayesian Image Analysis： Introduction1．The Bayesian Paradigm1．1 Warming up for Absolute Beginners1．2 Images and Observations1．3 Prior and Posterior Distributions1．4 Bayes Estimators2．Cleaning Dirty Pictures2．1 Boundaries and their Information Content2．2 Towards Piecewise Smoothing2．3 Filters， Smoothers， and Bayes Estimators2．4 Boundary Extraction2．5 Dependence on Hyperparameters3．Finite R，andom Fields3．1 Markov Random Fields3．2 Gibbs Fields and Potentials3．3 Potentials ContinuedPart Ⅱ．The Gibbs Sampler and Simulated Annealing4．Markov Chains： Limit Theorems4．1 Preliminaries4．2 The Contraction Coefficient4．3 Homogeneous Markov Chains4．4 Exact Sampling4．5 Inhomogeneous Markov Chains4．6 A Law of Large Numbers for Inhomogeneous Chains4．7 A Counterexample for the Law of Large Numbers5．Gibbsian Sampling and Annealing5．1 Sampling5．2 Simulated Annealing5．3 Discussion6．Cooling Schedules6．1 The ICM Algorithm6．2 Exact MAP Estimation Versus Fast Cooling6．3 Finite Time AnnealingPart Ⅲ．Variations of the Gibbs Sampler7．Gibbsian Sampling and Annealing Revisited7．1 A General Gibbs Sampler7．2 Sampling and Annealing under Constraints8．Partially Parallel Algorithms8．1 Synchronous Updating on Independent Sets8．2 The Swendson—Wang Algorithm9．Synchronous Algorithms9．1 Invariant Distributions and Convergence9．2 Support of the Limit Distribution9．3 Synchronous Algorithms and ReversibilityPart Ⅳ．Metropolis Algorithms and Spectral Methods10．Metropolis Algorithms10．1 Metropolis Sampling and Annealing10．2 Convergence Theorems10．3 Best Constants10．4 About Visiting Schemes10．5 Generalizations and Modifications10．6 The Metropolis Algorithm in Combinatorial Optimization11．The Spectral Gap and Convergence of Markov Chains11．1 Eigenvalues of Markov Kernels11．2 Geometric Convergence Rates12．Eigenvalues， Sampling， Variance Reduction12．1 Samplers and their Eigenvalues12．2 Variance Reduction12．3 Importance Sampling13．Continuous Time Processes13．1 Discrete State Space13．2 Continuous State SpacePart Ⅴ．Texture Analysis14．Partitioning14．1 How to Tell Textures Apart14．2 Bayesian Texture Segmentation14．3 Segmentation by a Boundary Model14．4 Julesz's Conjecture and Tw Point Processes15．Random Fields and Texture Models15．1 Neighbourhood Relations15．2 Random Field Texture Models15．3 Texture Synthesis16．Bayesian Texture Classification16．1 Contextual Classification16．2 Marginal Posterior Modes MethodsPart Ⅵ．Parameter Estimation17．Maximum Likelihood Estimation17．1 The Likelihood Function17．2 Objective Functions18．Consistency of Spatial ML Estimators18．1 Observation Windows and Specifications18．2 Pseudolikelihood Methods18．3 Large Deviations and Full Maximum Likelihood18．4 Partially Observed Data19．Computation of Full ML Estimators19．1 A Naive Algorithm19．2 Stochastic Optimization for the Full Likelihood19．3 Main Results19．4 Error Decomposition19．5 L2-EstimatesPart Ⅶ．Supplement20．A Glance at Neural Networks20．1 Boltzmann Machines20．2 A Learning Rule21．Three Applications21．1 Motion Analysis21．2 Tomographic Image Reconstruction21．3 Biological ShapePart Ⅷ．AppendixA．Simulation of Random VariablesA．1 Pseudorandom NumbersA．2 Discrete Random VariablesA．3 Special DistributionsB．Analytical ToolsB．1 Concave FunctionsB．2 Convergence of Descent AlgorithmsB．3 A Discrete Gronwall LemmaB．4 A Gradient SystemC．Physical Imaging SystemsD．The Software Package AntsInFieldsReferencesSymbolsIndex

第2版-图像分析.随机场和马尔可夫链蒙特卡罗方法 作者简介

Gerhard Winkler (G. 温克勒, 德国)是国际知名学者，在数学界享有盛誉。本书凝聚了作者多年科研和教学成果，适用于科研工作者、高校教师和研究生。