4.3 Multiscale Transform and Mathematical Morphology
4.4 Multiresolution Based on the Median Transform
4.5 Guided Numerical Experiments
4.6 Summary
5 The Ridgelet and Curvelet Transforms
5.1 Introduction
5.2 Background and Example
5.3 Ridgelets
5.4 Curvelets
5.5 Curvelets and Contrast Enhancement
5.6 Guided Numerical Experiments
5.7 Summary
6 Sparsity and Noise Removal
6.1 Introduction
6.2 Term-By-Term Nonlinear Denoising
6.3 Block Nonlinear Denoising
6.4 Beyond Additive Gaussian Noise
6.5 Poisson Noise and the Haar Transform
6.6 Poisson Noise with Low Counts
6.7 Guided Numerical Experiments
6.8 Summary
7 Linear Inverse Problems
7.1 Introduction
7.2 Sparsity-Regularized Linear Inverse Problems
7.3 Monotone Operator Splitting Framework
7.4 Selected Problems and Algorithms
7.5 Sparsity Penalty with Analysis Prior
7.6 Other Sparsity-Regularized Inverse Problems
7.7 General Discussion: Sparsity, Inverse Problems, and Iterative Thresholding
7.8 Guided Numerical Experiments
7.9 Summary
8 Morphological Diversity
8.1 Introduction
8.2 Dictionary and Fast Transformation
8.3 Combined Denoising
8.4 Combined Deconvolution
8.5 Morphological Component Analysis
8.6 Texture-Cartoon Separation
8.7 Inpainting
8.8 Guided Numerical Experiments
8.9 Summary
9 Sparse Blind Source Separation
9.1 Introduction
9.2 Independent Component Analysis
9.3 Sparsity and Multichannel Data
9.4 Morphological Diversity and Blind Source Separation
9.5 Illustrative Experiments
9.6 Guided Numerical Experiments
9.7 Summary
10 Multiscale Geometric Analysis on the Sphere
10.1 Introduction
10.2 Data on the Sphere
10.3 Orthogonal Haar Wavelets on the Sphere
10.4 Continuous Wavelets on the Sphere
10.5 Redundant Wavelet Transform on the Sphere with Exact Reconstruction
10.6 Curvelet Transform on the Sphere
10.7 Restoration and Decomposition on the Sphere
10.8 Applications
10.9 Guided Numerical Experiments
10.10 Summary
11 Compressed Sensing
11.1 Introduction
11.2 Incoherence and Sparsity
11.3 The Sensing Protocol
11.4 Stable Compressed Sensing
11.5 Designing Good Matrices: Random Sensing
11.6 Sensing with Redundant Dictionaries
11.7 Compressed Sensing in Space Science
11.8 Guided Numerical Experiments
11.9 Summary
References
List of Algorithms
Index
Nội dung
[...]... particular – and higher-order moments derived from the curvelet transform: wavelet scale 5, fourth-order moment curvelet band 1, second-order moment curvelet band 7, third- and fourth-order moments curvelet band 8, fourth-order moment curvelet band 11, fourth-order moment curvelet band 12, fourth-order moment curvelet band 16, fourth-order moment curvelet band 19, second- and fourth-order moments Our results... second-, third-, and fourth-order moments at each scale (hence variance, skewness, and kurtosis) So each image had 15 features For each of 19 bands resulting from the curvelet transform, we again determined the second-, third-, and fourth-order moments at each band (hence variance, skewness, and kurtosis) So each image had 57 features 13 14 Introduction to the World of Sparsity Figure 1.8 Sample images... imaging and signalprocessing in biology, medicine, and the life sciences generally; astronomy, physics, and the natural sciences; seismology and land use studies, as indicative subdomains from geology and geography in the earth sciences; materials science, metrology, and other areas of mechanical and civil engineering; image and video compression, analysis, and synthesis for movies and television; and. .. theory (Candes and Tao 2006; ` Donoho 2006a; Candes et al 2006b) Compressed sensing uses the prior knowledge that signals are sparse, whereas Shannon theory was designed for frequency band– limited signals By establishing a direct link between sampling and sparsity, compressed sensing has had a huge impact in many scientific fields such as coding and information theory, signal andimage acquisition and processing, ... transform xi Notation Functions and Signals f (t) f (t) or f (t1 , , td ) f [k] f [k] or f [k, l, ] f¯ fˆ f∗ H (z) lhs = O(rhs) lhs ∼ rhs 1{condition} L2 ( ) 2( ) (H) 0 continuous-time function, t ∈ R d-dimensional continuous-time function, t ∈ Rd discrete-time signal, k ∈ Z, or kth entry of a finite-dimensional vector d-dimensional discrete-time signal, k ∈ Zd time-reversed version of f as a function... Dirac dictionary (also known as standard unit vector basis or Kronecker basis), position scale for the wavelet dictionary, translation-duration-frequency for cosine packets, and position-scale-orientation for the curvelet dictionary in two dimensions In discrete-time, finite-length signal processing, a dictionary is viewed as an N × T matrix whose columns are the atoms, and the atoms are considered as... Figure 1.2 (left) SAR image of Gulf of Oman region and (right) resolutionscale information superimposed by the original image plus 100 times the resolution scale 3 image plus 20 times the resolution scale 4 image In Fig 1.3, the corresponding SeaWiFS image is shown The weighting used here for the right-hand image is the original image times 0.0005 plus the resolution scale 5 image In both cases, the... feature detection, andimage grading Applications range over Earth observation and astronomy, medicine, civil engineering and materials science, andimage databases generally 1.1 SPARSE REPRESENTATION 1.1.1 Introduction In the last decade, sparsity has emerged as one of the leading concepts in a wide range of signal- processing applications (restoration, feature extraction, source separation, and compression,... sparse if most of its entries are equal to zero, that is, if its support (x) = {1 ≤ i ≤ N | x[i] = 0} is of cardinality k N A k -sparse signal is a signal for which exactly k samples have a nonzero value If a signal is not sparse, it may be sparsified in an appropriate transform domain For instance, if x is a sine, it is clearly not sparse, but its Fourier transform is extremely sparse (actually, 1 -sparse) ... cited articles in statistics and signal processing, one finds works in the general area of what we cover in this book The methods discussed in this book are essential underpinnings of data analysis, of relevance to multimedia data processingand to image, video, and signalprocessing The methods discussed here feature very crucially in statistics, in mathematical methods, and in computational techniques . and index. ISBN 97 8-0 -5 2 1-1 191 3-9 (hardback) 1. Transformations (Mathematics) 2. Signal processing. 3. Image processing. 4. Sparse matrices. 5. Wavelets (Mathematics) I. Murtagh, Fionn. II. Fadili, Jalal,. in Publication data Starck, J L. (Jean-Luc), 1965– Sparse image and signal processing : wavelets, curvelets, morphological diversity / Jean-Luc Starck, Fionn Murtagh, Jalal Fadili. p. cm. Includes. h1" alt="" Sparse Image and Signal Processing: Wavelets, Curvelets, Morphological Diversity This book presents the state of the art in sparse and multiscale image and signal process- ing, covering