The transform and data compression handbook / edited by K.R. Rao and P.C. Yip.

Format:

Book

Contributor:

Rao, K. Ramamohan (Kamisetty Ramamohan)

Yip, P. C. (Pat C.)

Class of 1924 Book Fund.

Series:

Electrical engineering and signal processing series

The electrical engineering and signal processing series

Language:

English

Subjects (All):

Data transmission systems--Handbooks, manuals, etc.

Data transmission systems.

Data compression (Telecommunication).

Data compression (Telecommunication)--Handbooks, manuals, etc.

Genre:

Handbooks and manuals.

Physical Description:

xix, 388 pages : illustrations ; 24 cm.

Place of Publication:

Boca Raton, Fla. : CRC Press, [2001]

Summary:

Data compression is one of the main contributing factors in the explosive growth in information technology. Without it, a number of consumer and commercial products, such as DVD, videophone, digital camera, MP3, video-streaming and wireless PCS, would have been virtually impossible. Transforming the data to a frequency or other domain enables even more efficient compression. By illustrating this intimate link, The Transform and Data Compression Handbook serves as a much-needed handbook for a wide range of researchers and engineers. The authors describe various discrete transforms and their applications in different disciplines. They cover techniques, such as adaptive quantization and entropy coding, that result in significant reduction in bit rates when applied to the transform coefficients. With clear and concise presentations of the ideas and concepts, as well as detailed descriptions of the algorithms, the authors provide important insight into the applications and their limitations. Data compression is an essential step towards the efficient storage and transmission of information. The Transform and Data Compression Handbook provides a wealth of information regarding different discrete transforms and demonstrates their power and practicality in data compression.

Contents:

1 Karhunen-Loeve Transform 1

1.2 Data Decorrelation 2

1.2.1 Calculation of the KLT 9

1.3 Performance of Transforms 11

1.3.1 Information Theory 11

1.3.2 Quantization 13

1.3.3 Truncation Error 13

1.3.4 Block Size 15

1.4.1 Calculation of KLT 17

1.4.2 Quantization and Encoding 18

1.4.3 Generalization 22

1.4.4 Markov-1 Solution 24

1.4.5 Medical Imaging 25

1.4.6 Color Images 28

2 The Discrete Fourier Transform 37

2.2 The DFT Matrix 39

2.3 An Example 40

2.4 DFT Frequency Analysis 41

2.5 Selected Properties of the DFT 45

2.5.1 Symmetry Properties 47

2.6 Real-Valued DFT-Based Transforms 49

2.7 The Fast Fourier Transform 55

2.8 The DFT in Coding Applications 58

2.9 The DFT and Filter Banks 60

2.9.1 Cosine-Modulated Filter Banks 63

2.9.2 Complex DFT-Based Filter Banks 66

2.11 FFT Web sites 72

3 Comparametric Transforms for Transmitting Eye Tap Video with Picture Transfer Protocol (PTP) 79

3.1 Introduction: Wearable Cybernetics 79

3.1.1 Historical Overview of WearComp 80

3.1.2 Eye Tap Video 80

3.2 The Edgertonian Image Sequence 81

3.2.1 Edgertonian versus Nyquist Thinking 81

3.2.2 Frames versus Rows, Columns, and Pixels 82

3.3 Picture Transfer Protocol (PTP) 83

3.4 Best Case Imaging and Fear of Functionality 84

3.5 Comparametric Image Sequence Analysis 88

3.5.1 Camera, Eye, or Head Motion: Common Assumptions and Terminology 91

3.5.2 VideoOrbits 92

3.6 Framework: Comparameter Estimation and Optical Flow 94

3.6.1 Feature-Based Methods 94

3.6.2 Featureless Methods Based on Generalized Cross-Correlation 95

3.6.3 Featureless Methods Based on Spatio-Temporal Derivatives 96

3.7 Multiscale Projective Flow Comparameter Estimation 99

3.7.1 Four Point Method for Relating Approximate Model to Exact Model 101

3.7.2 Overview of the New Projective Flow Algorithm 102

3.7.3 Multiscale Repetitive Implementation 103

3.7.4 Exploiting Commutativity for Parameter Estimation 104

3.8 Performance/Applications 106

3.8.1 A Paradigm Reversal in Resolution Enhancement 106

3.8.2 Increasing Resolution in the "Pixel Sense" 107

4 Discrete Cosine and Sine Transforms 117

4.2 The Family of DCTs and DSTs 118

4.2.1 Definitions of DCTs and DSTs 118

4.2.2 Mathematical Properties 119

4.2.3 Relations to the KLT 121

4.3 A Unified Fast Computation of DCTs and DSTs 122

4.3.1 Definitions of Even-Odd Matrices 123

4.3.2 DCT-II/DST-II and DCT-III/DST-III Computation 124

4.3.3 DCT-I and DST-I Computation 129

4.3.4 DCT-IV/DST-IV Computation 131

4.3.5 Implementation of the Unified Fast Computation of DCTs and DSTs 134

4.4 The 2-D DCT/DST Universal Computational Structure 146

4.4.1 The Fast Direct 2-D DCT/DST Computation 146

4.4.2 Implementation of the Direct 2-D DCT/DST Computation 152

4.5 DCT and Data Compression 161

4.5.1 DCT-Based Image Compression/Decompression 162

4.5.2 Data Structures for Compression/Decompression 166

4.5.3 Setting the Quantization Table 168

4.5.4 Standard Huffman Coding/Decoding Tables 170

4.5.5 Compression of One Sub-Image Block 172

4.5.6 Decompression of One Sub-Image Block 179

4.5.7 Image Compression/Decompression 184

4.5.8 Compression of Color Images 186

4.5.9 Results of Image Compression 188

5 Lapped Transforms for Image Compression 197

5.1.2 Brief History 198

5.1.3 Block Transforms 199

5.1.4 Factorization of Discrete Transforms 200

5.1.5 Discrete MIMO Linear Systems 201

5.1.6 Block Transform as a MIMO System 203

5.2 Lapped Transforms 204

5.2.1 Orthogonal Lapped Transforms 204

5.2.2 Nonorthogonal Lapped Transforms 210

5.3 LTs as MIMO Systems 210

5.4 Factorization of Lapped Transforms 213

5.5 Hierarchical Connection of LTs: An Introduction 215

5.5.1 Time-Frequency Diagram 215

5.5.2 Tree-Structured Hierarchical Lapped Transforms 217

5.5.3 Variable-Length LTs 219

5.6 Practical Symmetric LTs 222

5.6.1 The Lapped Orthogonal Transform: LOT 222

5.6.2 The Lapped Bi-Orthogonal Transform: LBT 223

5.6.3 The Generalized LOT: GenLOT 226

5.6.4 The General Factorization: GLBT 230

5.7 The Fast Lapped Transform: FLT 233

5.8 Modulated LTs 236

5.9 Finite-Length Signals 240

5.9.1 Overall Transform 241

5.9.2 Recovering Distorted Samples 243

5.9.3 Symmetric Extensions 244

5.10 Design Issues for Compression 246

5.11 Transform-Based Image Compression Systems 248

5.11.1 JPEG 249

5.11.2 Embedded Zerotree Coding 250

5.11.3 Other Coders 252

5.12 Performance Analysis 253

5.12.1 JPEG 253

5.12.2 Embedded Zerotree Coding 255

6 Wavelet-Based Image Compression 267

6.2 Dyadic Wavelet Transform 268

6.2.1 Two-Channel Perfect-Reconstruction Filter Bank 270

6.2.2 Dyadic Wavelet Transform, Multiresolution Representation 272

6.2.3 Wavelet Smoothness 273

6.3 Wavelet-Based Image Compression 274

6.3.1 Lossy Compression 274

6.3.2 EZW Algorithm 278

6.3.3 SPIHT Algorithm 285

6.3.4 WDR Algorithm 294

6.3.5 ASWDR Algorithm 299

6.3.6 Lossless Compression 305

6.3.7 Color Images 305

6.3.8 Other Compression Algorithms 306

6.3.9 Ringing Artifacts and Postprocessing Algorithms 306

7 Fractal-Based Image and Video Compression 313

7.2 Basic Properties of Fractals and Image Compression 314

7.3 Contractive Affine Transforms, Iterated Function Systems, and Image Generation 316

7.4 Image Compression Directly Based on the IFS Theory 318

7.5 Image Compression Based on IFS Library 321

7.6 Image Compression Based on Partitioned IFS 322

7.6.1 Image Partitions 323

7.6.2 Distortion Measure 323

7.6.3 A Class of Discrete Image Transformations 324

7.6.4 Encoding and Decoding Procedures 325

7.6.5 Experimental Results 326

7.7 Image Coding Using Quadtree Partitioned IFS (QPIFS) 326

7.7.1 RMS Tolerance Selection 328

7.7.2 A Compact Storage Scheme 329

7.7.3 Experimental Results 331

7.8 Image Coding by Exploiting Scalability of Fractals 333

7.8.1 Image Spatial Sub-Sampling 334

7.8.2 Decoding to a Larger Image 334

7.8.3 Experimental Results 334

7.9 Video Sequence Compression using Quadtree PIFS 336

7.9.1 Definitions of Types of Range Blocks 336

7.9.2 Encoding and Decoding Processes 338

7.9.3 Storage Requirements 340

7.9.4 Experimental Results 340

7.10 Other Fractal-Based Image Compression Techniques 341

7.10.1 Segmentation-Based Coding Using Fractal Dimension 341

7.10.2 Yardstick Coding 342

8 Compression of Wavelet Transform Coefficients 347

8.2 Embedded Coefficient Coding 353

8.3 Statistical Context Modeling of Embedded Bit Stream 357

8.4 Context Dilution Problem 359

8.5 Context Formation 360

8.6 Context Quantization 362

8.7 Optimization of Context Quantization 365

8.8 Dynamic Programming for Minimum Conditional Entropy 367

8.9 Fast Algorithms for High-Order Context Modeling 369

8.9.1 Context Formation via Convolution 370

8.9.2 Shared Modeling Context for Signs and Textures 371

8.10 Experimental Results 373

8.10.1 Lossy Case 373

8.10.2 Lossless Case 374.

Notes:

Includes bibliographical references and index.

Local Notes:

Acquired for the Penn Libraries with assistance from the Class of 1924 Book Fund.

ISBN:

0849336929

OCLC:

44468841