首页--工业技术论文--自动化技术、计算机技术论文--计算技术、计算机技术论文--计算机的应用论文--信息处理(信息加工)论文--模式识别与装置论文

面向低层视觉的稀疏低秩模型理论与方法

ABSTRACT第5-7页
摘要第8-15页
List of Symbols第15-16页
List of Abbreviations第16-23页
Chapter 1 Introduction第23-35页
    1.1 Sparse and low rank in theory第23-29页
        1.1.1 Exact representation via atom decomposition第23-24页
        1.1.2 From exact to approximate solution第24-25页
        1.1.3 Sparse and low rank via l_1-norm第25-27页
        1.1.4 Advanced sparse and low rank through lp quasi norm第27-28页
        1.1.5 An MAP perspective第28-29页
    1.2 Sparse and low rank in low level vision第29-31页
        1.2.1 Sparse representation of the image patches第30页
        1.2.2 From sparse to low rank model第30-31页
    1.3 Contributions and organization第31-35页
Chapter 2 Preliminaries第35-39页
    2.1 Prior arts第35-36页
        2.1.1 The gauge function and atomic norm第35页
        2.1.2 The rank and SVD of matrix第35-36页
        2.1.3 Nuclear norm-a variational formulation第36页
    2.2 Related optimization algorithms第36-39页
        2.2.1 Some basic concepts in optimization第37页
        2.2.2 Advanced optimization algorithms第37-39页
Chapter 3 Rank constrained nuclear norm minimization for image denoising第39-57页
    3.1 Introduction第39-41页
    3.2 Rank constrained nuclear norm minimization第41-45页
    3.3 Optimization-SOR method for RNNM第45-49页
        3.3.1 Convergence analysis第46-49页
        3.3.2 Complexity analysis第49页
    3.4 RNNM for image denoising第49-51页
    3.5 Experiments第51-57页
        3.5.1 Comparison of the computational efficiency第51-52页
        3.5.2 Denoising results on tested images第52-57页
Chapter 4 Generalized unitarily invariant gauge function for low level vision第57-87页
    4.1 Introduction第57-59页
    4.2 Related work第59-61页
        4.2.1 Gauge function第59-60页
        4.2.2 Alternating minimization of bi-linear model第60-61页
    4.3 Generalized unitarily invariant gauge (GUIG) function第61-68页
        4.3.1 The proposed GUIG function第61-64页
        4.3.2 Bilinear representation of the GUIG function第64-66页
        4.3.3 GUIG regularization第66-68页
    4.4 GUIG regularized fast matrix recovery第68-75页
        4.4.1 Matrix completion第68-71页
        4.4.2 Robust principle component analysis第71-73页
        4.4.3 Convergence analysis第73-75页
        4.4.4 Complexity analysis第75页
    4.5 Experimental results第75-87页
        4.5.1 Matrix completion第76-83页
        4.5.2 RPCA第83-87页
Chapter 5 Online Schatten-p quasi norm minimization for robust principle compo-nent analysis第87-107页
    5.1 Introduction第87-88页
    5.2 Related work第88-89页
    5.3 Problem formulation第89-92页
        5.3.1 Bi-linear representation of the Schatten-p quasi norm regularized R-PCA第90-91页
        5.3.2 Online reformulation via column separation第91-92页
    5.4 Stochastic optimization第92-95页
        5.4.1 Estimate the coefficient and the noise第93-94页
        5.4.2 Update the basis matrix第94-95页
    5.5 Theoretical analysis第95-98页
        5.5.1 Convergence analysis第95-98页
        5.5.2 Memory and computational cost第98页
    5.6 Experiments第98-107页
        5.6.1 Subspace recovery第99-101页
        5.6.2 Video background subtraction第101-107页
Chapter 6 Bayesian inference for adaptive low rank and sparse estimation第107-129页
    6.1 Introduction第107-108页
    6.2 Related work第108-111页
        6.2.1 Applications of the low rank matrix approximation第108-109页
        6.2.2 NNM to generalized low rank approximation第109-110页
        6.2.3 Statistical interpretation第110-111页
    6.3 Bayesian inference for adaptive low rank matrix approximation第111-116页
        6.3.1 Bayesian inference via full MAP第111-113页
        6.3.2 Influence of the singular values σ_(yi)第113页
        6.3.3 Global minimum of the non-convex optimization problem第113-116页
    6.4 Low rank approximation with sparse outliers (ARLLRE)第116-120页
        6.4.1 The ARLLRE model第116-118页
        6.4.2 Convergence analysis第118-120页
    6.5 Experiments第120-129页
        6.5.1 ARLLR for image denoising第122-124页
        6.5.2 ARLLRE for robust principle component analysis第124-129页
Chapter 7 Conclusions第129-131页
References第131-143页
Acknowledgements第143-145页
Curriculum Vitae第145-146页

论文共146页,点击 下载论文
上一篇:行走压能PVDF压电集能器的构成与性能表征
下一篇:基于计算机视觉的手势识别及人机交互技术的应用研究