Muutke küpsiste eelistusi

Super-Resolution Imaging [Kõva köide]

4.00/5 (4 hinnangut Goodreads-ist)
Edited by (University of California, Santa Cruz, USA)
  • Formaat: Hardback, 496 pages, kõrgus x laius: 234x156 mm, kaal: 839 g, 13 Tables, black and white; 163 Illustrations, black and white
  • Sari: Digital Imaging and Computer Vision
  • Ilmumisaeg: 28-Sep-2010
  • Kirjastus: CRC Press Inc
  • ISBN-10: 1439819300
  • ISBN-13: 9781439819302
Teised raamatud teemal:
Super-Resolution ImagingSuurem pilt
  • Formaat: Hardback, 496 pages, kõrgus x laius: 234x156 mm, kaal: 839 g, 13 Tables, black and white; 163 Illustrations, black and white
  • Sari: Digital Imaging and Computer Vision
  • Ilmumisaeg: 28-Sep-2010
  • Kirjastus: CRC Press Inc
  • ISBN-10: 1439819300
  • ISBN-13: 9781439819302
Teised raamatud teemal:
Püsilink: https://www.kriso.ee/db/9781439819302.html
Märksõnad:
With the exponential increase in computing power and broad proliferation of digital cameras, super-resolution imaging is poised to become the next "killer app." The growing interest in this technology has manifested itself in an explosion of literature on the subject. Super-Resolution Imaging consolidates key recent research contributions from eminent scholars and practitioners in this area and serves as a starting point for exploration into the state of the art in the field. It describes the latest in both theoretical and practical aspects of direct relevance to academia and industry, providing a base of understanding for future progress.

Features downloadable tools to supplement material found in the book

Recent advances in camera sensor technology have led to an increasingly larger number of pixels being crammed into ever-smaller spaces. This has resulted in an overall decline in the visual quality of recorded content, necessitating improvement of images through the use of post-processing. Providing a snapshot of the cutting edge in super-resolution imaging, this book focuses on methods and techniques to improve images and video beyond the capabilities of the sensors that acquired them. It covers:











History and future directions of super-resolution imaging





Locally adaptive processing methods versus globally optimal methods





Modern techniques for motion estimation





How to integrate robustness





Bayesian statistical approaches





Learning-based methods





Applications in remote sensing and medicine





Practical implementations and commercial products based on super-resolution

The book concludes by concentrating on multidisciplinary applications of super-resolution for a variety of fields. It covers a wide range of super-resolution imaging implementation techniques, including variational, feature-based, multi-channel, learning-based, locally adaptive, and nonparametric methods. This versatile book can be used as the basis for short courses for engineers and scientists, or as part of graduate-level courses in image processing.
Preface xv
1 Image Super-Resolution: Historical Overview and Future Challenges
1(34)
Jianchao Yang
Thomas Huang
1.1 Introduction to Super-Resolution
1(4)
1.2 Notations
5(1)
1.3 Techniques for Super-Resolution
5(15)
1.3.1 Image Observation Model
5(2)
1.3.2 Super-Resolution in the Frequency Domain
7(1)
1.3.3 Interpolation-Restoration: Non-Iterative Approaches
8(1)
1.3.4 Statistical Approaches
9(2)
1.3.4.1 Maximum Likelihood
11(1)
1.3.4.2 Maximum a Posteriori
12(1)
1.3.4.3 Joint MAP Restoration
13(1)
1.3.4.4 Bayesian Treatments
14(1)
1.3.5 Example-Based Approaches
15(3)
1.3.6 Set Theoretic Restoration
18(2)
1.4 Challenge Issues for Super-Resolution
20(4)
1.4.1 Image Registration
20(1)
1.4.2 Computation Efficiency
21(1)
1.4.3 Robustness Aspects
22(1)
1.4.4 Performance Limits
23(1)
Bibliography
24(11)
2 Super-Resolution Using Adaptive Wiener Filters
35(28)
Russell C. Hardie
2.1 Introduction
36(2)
2.2 Observation Model
38(9)
2.2.1 Image Formation Model
38(3)
2.2.2 Image Motion Model
41(1)
2.2.3 Image Registration
42(2)
2.2.4 System Point-Spread Function
44(3)
2.3 AWF SR Algorithms
47(4)
2.4 Experimental Results
51(6)
2.4.1 SR Results for Simulated Data
51(2)
2.4.2 SR Results for Infrared Video Data
53(4)
2.5 Conclusions
57(1)
2.6 Acknowledgments
58(1)
Bibliography
58(5)
3 Locally Adaptive Kernel Regression for Space-Time Super-Resolution
63(34)
Hiroyuki Takeda
Peyman Milanfar
3.1 Introduction
64(3)
3.2 Adaptive Kernel Regression
67(16)
3.2.1 Classic Kernel Regression in 2-D
67(3)
3.2.2 Steering Kernel Regression in 2-D
70(2)
3.2.3 Space-Time (3-D) Steering Kernel Regression
72(6)
3.2.4 Kernel Regression with Rough Motion Compensation
78(2)
3.2.5 Implementation and Iterative Refinement
80(3)
3.3 Examples
83(7)
3.3.1 Spatial Upscaling Examples
83(3)
3.3.2 Spatiotemporal Upscaling Examples
86(4)
3.4 Conclusion
90(1)
3.5 Appendix
91(1)
3.5.1 Steering Kernel Parameters
91(1)
3.5.2 The Choice of the Regression Parameters
91(1)
3.5.3 Deblurring
92(1)
Bibliography
92(5)
4 Super-Resolution with Probabilistic Motion Estimation
97(26)
Matan Protter
Michael Elad
4.1 Introduction
98(1)
4.2 Classic Super-Resolution: Background
99(2)
4.3 The Proposed Algorithm
101(7)
4.3.1 The New Formulation
101(2)
4.3.2 Separating the Blur Treatment
103(1)
4.3.3 The Algorithm: A Matrix-Vector Version
104(1)
4.3.4 The Algorithm: A Pixel-Wise Version
104(2)
4.3.5 Computing the Weights
106(1)
4.3.6 Other Resampling Tasks
107(1)
4.4 Experimental Validation
108(10)
4.4.1 Experimental Results
108(8)
4.4.2 Computational Complexity
116(2)
4.5 Summary
118(1)
Bibliography
119(4)
5 Spatially Adaptive Filtering as Regularization in Inverse Imaging: Compressive Sensing, Super-Resolution, and Upsampling
123(32)
Aram Danielyan
Alessandro Foi
Vladimir Katkovnik
Karen Egiazarian
5.1 Introduction
124(1)
5.2 Iterative Filtering as Regularization
125(4)
5.2.1 Spectral Decomposition of the Operator
126(1)
5.2.2 Nonlocal Transform Domain Filtering
126(3)
5.3 Compressed Sensing
129(8)
5.3.1 Observation Model and Notation
130(1)
5.3.2 Iterative Algorithm with Stochastic Approximation
130(2)
5.3.2.1 Comments on the Algorithm
132(1)
5.3.3 Experiments
133(1)
5.3.3.1 Radon Inversion from Sparse Projections
134(1)
5.3.3.2 Limited-Angle Tomography
134(1)
5.3.3.3 Reconstruction from Low-Frequency Data
134(3)
5.4 Super-Resolution
137(13)
5.4.1 Spectral Decomposition for the Super-Resolution Problem
139(1)
5.4.2 Observation Model
140(1)
5.4.3 Scaling Family of Transforms
140(2)
5.4.4 Multistage Iterative Reconstruction
142(1)
5.4.5 Experiments
143(1)
5.4.5.1 Implementation Details
144(1)
5.4.5.2 Super-Resolution
144(1)
5.4.5.3 Image Upsampling
145(5)
5.5 Conclusions
150(1)
Bibliography
150(5)
6 Registration for Super-Resolution: Theory, Algorithms, and Applications in Image and Mobile Video Enhancement
155(32)
Patrick Vandewalle
Luciano Sbaiz
Martin Vetterli
6.1 Camera Model
157(4)
6.2 What is Resolution?
161(1)
6.3 Super-Resolution as a Multichannel Sampling Problem
162(4)
6.3.1 Fourier Series
163(3)
6.4 Registration of Totally Aliased Signals
166(2)
6.4.1 Variable Projection Method
166(1)
6.4.2 Frequency Analysis Method
167(1)
6.4.3 Results
168(1)
6.5 Registration of Partially Aliased Signals
168(15)
6.5.1 Super-Resolution Using Frequency Domain Registration
168(1)
6.5.1.1 Image Registration
168(2)
6.5.1.2 Image Reconstruction
170(1)
6.5.1.3 Results
170(6)
6.5.2 Super-Resolution from Low-Quality Videos
176(1)
6.5.2.1 Motion Model
176(1)
6.5.2.2 Image Registration
177(2)
6.5.2.3 Image Reconstruction
179(2)
6.5.2.4 Results on Video Sequences
181(2)
6.6 Conclusions
183(1)
Bibliography
184(3)
7 Towards Super-Resolution in the Presence of Spatially Varying Blur
187(32)
Michal Sorel
Filip Sroubek
Jan Flusser
7.1 Introduction
188(6)
7.1.1 Representation of Spatially Varying PSF
189(1)
7.1.2 General Model of Resolution Loss
189(2)
7.1.3 Bayesian View of Solution
191(3)
7.2 Defocus and Optical Aberrations
194(8)
7.2.1 Geometrical Optics
195(2)
7.2.2 Approximation of PSF by 2D Gaussian Function
197(1)
7.2.3 General Form of PSF for Axially-Symmetric Optical Systems
197(1)
7.2.4 Diffraction
198(3)
7.2.5 Summary
201(1)
7.3 Camera Motion Blur
202(2)
7.3.1 Rotation
202(1)
7.3.2 No Rotation
203(1)
7.4 Scene Motion
204(1)
7.5 Algorithms
204(10)
7.5.1 Super-Resolution of a Scene with Local Motion
205(2)
7.5.2 Smoothly Changing Blur
207(3)
7.5.3 Depth-Dependent Blur
210(4)
7.6 Conclusion
214(1)
7.7 Acknowledgments
215(1)
Bibliography
215(4)
8 Toward Robust Reconstruction-Based Super-Resolution
219(28)
Masayuki Tanaka
Masatoshi Okutomi
8.1 Introduction
220(1)
8.2 Overviews
221(4)
8.2.1 Super-Resolution Reconstruction
221(1)
8.2.2 Robust SR Reconstruction
222(2)
8.2.3 Robust Registration
224(1)
8.3 Robust SR Reconstruction with Pixel Selection
225(7)
8.3.1 Displacement and Similarity Measure
225(2)
8.3.2 Proposed Pixel Selection Algorithm
227(1)
8.3.2.1 Pixel Selection Based on Similarity Measure and Displacement Estimation
227(2)
8.3.3 Luminance Correction
229(1)
8.3.4 Experiments
230(2)
8.4 Robust Super-Resolution Using MPEG Motion Vectors
232(5)
8.4.1 Registration Using MPEG Motion Vectors
232(1)
8.4.2 Experiments of Robust SR Reconstruction
233(4)
8.5 Robust Registration for Super-Resolution
237(7)
8.5.1 Proposed Multiple Motion Estimation
238(1)
8.5.1.1 Motion Estimation and Region Extraction for a Single Object
239(1)
8.5.1.2 Multiple Motion Estimation
240(1)
8.5.2 Super-Resolution for Multiple Motions
241(1)
8.5.3 Experiments
241(3)
8.6 Conclusions
244(1)
Bibliography
244(3)
9 Multiframe Super-Resolution from a Bayesian Perspective
247(38)
Lyndsey Pickup
Stephen Roberts
Andrew Zisserman
David Capel
9.1 The Generative Model
248(10)
9.1.1 Considerations in the Forward Model
249(2)
9.1.2 A Probabilistic Setting
251(1)
9.1.2.1 The Maximum Likelihood Solution
251(1)
9.1.2.2 The ML Solution in Practice
252(2)
9.1.2.3 The Maximum a Posteriori Solution
254(1)
9.1.3 Selected Priors Used in MAP Super-Resolution
255(3)
9.2 Where Super-Resolution Algorithms Go Wrong
258(5)
9.2.1 Point-Spread Function Example
259(2)
9.2.2 Photometric Registration Example
261(1)
9.2.3 Geometric Registration Example
262(1)
9.3 Simultaneous Super-Resolution
263(10)
9.3.1 Super-Resolution with Registration
264(1)
9.3.2 Learning Prior Strength Parameters from Data
265(1)
9.3.3 Scaling and Convergence
266(1)
9.3.4 Initialization
267(2)
9.3.5 Evaluation on Synthetic Data
269(2)
9.3.6 Experiments on Real Data
271(2)
9.4 Bayesian Marginalization
273(9)
9.4.1 Marginalizing over Registration Parameters
274(3)
9.4.2 Marginalizing over the High-Resolution Image
277(1)
9.4.3 Implementation Notes
278(1)
9.4.4 Experimental Evaluation
279(3)
9.4.5 Discussion
282(1)
9.5 Concluding Remarks
282(1)
Bibliography
283(2)
10 Variational Bayesian Super-Resolution Reconstruction
285(30)
S. Derin Babacan
Rafael Molina
Aggelos K. Katsaggelos
10.1 Introduction
285(3)
10.2 Problem Formulation
288(1)
10.3 Bayesian Framework for Super-Resolution
288(5)
10.3.1 Observation Models
289(1)
10.3.2 Image Models
290(1)
10.3.3 Blur Models
291(1)
10.3.4 Motion (Registration) Models
292(1)
10.3.5 Hyperpriors on the Hyperparameters
292(1)
10.4 Bayesian Inference
293(3)
10.5 Variational Bayesian Inference Using TV Image Priors
296(5)
10.5.1 Estimation of the HR Image Distribution
297(1)
10.5.2 Estimation of the Hyperparameter Distributions
298(3)
10.6 Experiments
301(4)
10.7 Estimation of Motion and Blur
305(3)
10.8 Conclusions
308(1)
10.9 Acknowledgments
309(1)
Bibliography
309(6)
11 Pattern Recognition Techniques for Image Super-Resolution
315(40)
Karl Ni
Truong Q. Nguyen
11.1 Introduction
316(2)
11.2 Nearest Neighbor Super-Resolution
318(8)
11.2.1 k-Nearest Neighbor
320(1)
11.2.2 k-Nearest Neighbor Regression
321(2)
11.2.3 Adaptive k-NN for Super-Resolution
323(2)
11.2.4 Heuristics for Insufficient Training in Adaptive k-NN Regression
325(1)
11.3 Markov Random Fields and Approximations
326(3)
11.4 Kernel Machines for Image Super-Resolution
329(9)
11.4.1 Support Vector Regression
330(2)
11.4.2 Inductively Learning the Kernel Matrix for Regression
332(3)
11.4.3 The Quadratically Constrained Quadratic Programming Problem
335(2)
11.4.4 Applications to Super-Resolution
337(1)
11.5 Multiple Learners and Multiple Regressions
338(8)
11.5.1 Neural Networks and Super-Resolution
339(1)
11.5.2 Unsupervised Clustering
340(2)
11.5.3 Supervised Clustering
342(1)
11.5.4 Integrating Regression
343(3)
11.6 Design Considerations and Examples
346(2)
11.7 Remarks
348(1)
11.8 Glossary
349(1)
Bibliography
349(6)
12 Super-Resolution Reconstruction of Multichannel Images
355(28)
Osman G. Sezer
Yucel Altunbasak
12.1 Introduction
356(2)
12.2 Notation
358(2)
12.3 Image Acquisition Model
360(6)
12.3.1 Motion Compensation
361(1)
12.3.2 Spatial Filtering
362(1)
12.3.3 Spectral Filtering
363(2)
12.3.4 Multichannel Observation Model
365(1)
12.4 Subspace Representation
366(3)
12.4.1 Blind Source Separation
367(1)
12.4.2 Observation Model with BSS
368(1)
12.5 Reconstruction Algorithm
369(4)
12.5.1 The Subspace Observation Model
370(1)
12.5.2 POCS with Outliers of Residual
371(1)
12.5.3 POCS with Variance of Residual
372(1)
12.6 Experiments and Discussions
373(4)
12.6.1 Spectral Subspace
374(1)
12.6.2 Robustness against Noise
375(1)
12.6.3 Simultaneous Spatial and Spectral Super-Resolution
376(1)
12.7 Conclusion
377(3)
Bibliography
380(3)
13 New Applications of Super-Resolution in Medical Imaging
383(30)
M. Dirk Robinson
Stephanie J. Chiu
Cynthia A. Toth
Joseph A. Izatt
Joseph Y. Lo
Sina Farsiu
13.1 Introduction
384(1)
13.2 The Super-Resolution Framework
385(3)
13.2.1 Image Capture Model
385(2)
13.2.2 Super-Resolution Estimation Framework
387(1)
13.3 New Medical Imaging Applications
388(17)
13.3.1 Super-Resolution in Low Radiation Digital X-Ray Mammography
389(2)
13.3.1.1 Multiframe Shift Estimation
391(2)
13.3.1.2 Multiframe forWaRD Deconvolution and Denoising
393(1)
13.3.1.3 Experimental X-Ray Results
394(3)
13.3.2 Super-Resolution in Optical Coherence Tomography
397(2)
13.3.2.1 Proposed Method: Sparse Repeated Imaging
399(2)
13.3.2.2 Multiframe Joint Registration
401(2)
13.3.2.3 Experimental Results
403(2)
13.4 Conclusion
405(1)
13.5 Acknowledgments
406(1)
Bibliography
407(6)
14 Practicing Super-Resolution: What Have We Learned?
413(36)
Nikola Bozinovic
14.1 Abstract
414(1)
14.2 Introduction
414(6)
14.2.1 Video Quality Trends
415(1)
14.2.2 The Need for Postprocessing
415(2)
14.2.3 Why is Super-Resolution not Used More?
417(1)
14.2.4 Automation versus User Interaction
418(1)
14.2.5 Modeling Motion for Super-Resolution
418(1)
14.2.6 Performance Issues
419(1)
14.2.7 Relationship to Existing Standards
419(1)
14.3 MotionDSP: History and Concepts
420(2)
14.3.1 Design Decisions
421(1)
14.4 Markets and Applications
422(5)
14.4.1 Forensic and Real-Time Markets: MotionDSP's Ikena
423(2)
14.4.2 Consumers: MotionDSP's vReveal
425(2)
14.5 Technology
427(4)
14.5.1 Robust Parametric Motion Estimation
428(3)
14.6 Results
431(6)
14.6.1 Mobile and Digital Still Camera Video
432(1)
14.6.2 DV and HD Video
433(1)
14.6.3 Handling Complex Motion
433(2)
14.6.4 Practical Limits of Super-Resolution
435(2)
14.7 Lessons Learned
437(7)
14.8 Conclusions
444(1)
Bibliography
444(5)
Index 449
Peyman Milanfar is Professor of Electrical Engineering at the University of California, Santa Cruz. He received a B.S. degree in Electrical Engineering/Mathematics from the University of California, Berkeley, and the Ph.D. degree in Electrical Engineering from the Massachusetts Institute of Technology. Prior to coming to UCSC, he was at SRI (formerly Stanford Research Institute) and served as a Consulting Professor of computer science at Stanford. In 2005 he founded MotionDSP, Inc., to bring state-of-art video enhancement technology to consumer and forensic markets. He is a Fellow of the IEEE for contributions to Inverse Problems and Super-resolution in Imaging.