.

0 (hodnocen0 x )

(2) Půjčeno:2x 

BK

Fourth edition

Australia : Cengage Learning, [2015]

xxxv, 870 stran : ilustrace (některé barevné) ; 24 cm

ISBN 978-1-133-59369-0 (brožováno)

"International edition"--Obálka

Obsahuje bibliografie a rejstřík

001463438

List of algorithms xxi // Preface xxv // Possible course outlines xxxi // 1 Introduction 1 // 1.1 Motivation 1 // 1.2 Why is computer vision difficult? 3 // 1.3 Image representation and image analysis tasks 5 // 1.4 Summary 9 // 1.5 Exercises 10 // 1.6 References 10 // 2 The image, its representations and properties 11 // 2.1 Image representations, a few concepts 11 // 2.2 Image digitization 14 // 2.2.1 Sampling 14 // 2.2.2 Quantization 15 // 2.3 Digital image properties 16 // 2.3.1 Metric and topological properties of digital images 16 // 2.3.2 Histograms 23 // 2.3.3 Entropy 24 // 2.3.4 Visual perception of the image 25 // 2.3.5 Image quality 27 // 2.3.6 Noise in images 28 // 2.4 Color images 30 // 2.4.1 Physics of color 30 // 2.4.2 Color perceived by humans 32 // 2.4.3 Color spaces 36 // 2.4.4 Palette images 38 // 2.4.5 Color constancy 39 // 2.5 Cameras: An overview 40 // 2.5.1 Photosensitive sensors 40 // 2.5.2 A monochromatic camera 42 // 2.5.3 A color camera 44 // 2.6 Summary 45 // 2.7 Exercises 46 // 2.8 References 48 // 3 The image, its mathematical and physical background 50 // 3.1 Overview 50 // 3.1.1 Linearity 50 // 3.1.2 The Dirac distribution and convolution 51 // 3.2 Linear integral transforms 52 // 3.2.1 Images as linear systems 53 // 3.2.2 Introduction to linear integral transforms 53 // 3.2.3 ID Fourier transform 54 // 3.2.4 2D Fourier transform 59 // 3.2.5 Sampling and the Shannon constraint 62 // 3.2.6 Discrete cosine transform 65 // 3.2.7 Wavelet transform 66 // 3.2.8 Eigen-analysis 72 // 3.2.9 Singular value decomposition 73 // 3.2.10 Principal component analysis 74 // 3.2.11 Radon transform 77 // 3.2.12 Other orthogonal image transforms 78 // 3.3 Images as stochastic processes 79 // 3.4 Image formation physics 82 // 3.4.1 Images as radiometric measurements 82 // 3.4.2 Image capture and geometric optics 83 //

3.4.3 Lens aberrations and radial distortion 86 // 3.4.4 Image capture from a radiometric point of view 89 // 3.4.5 Surface reflectance 92 // 3.5 Summary 95 // 3.6 Exercises 97 // 3.7 References 98 // 4 Data structures for image analysis 100 // 4.1 Levels of image data representation 100 // 4.2 Traditional image data structures 101 // 4.2.1 Matrices 101 // 4.2.2 Chains 104 // 4.2.3 Topological data structures 106 // 4.2.4 Relational structures 107 // 4.3 Hierarchical data structures 108 // 4.3.1 Pyramids 108 // 4.3.2 Quadtrees 109 // 4.3.3 Other pyramidal structures 111 // 4.4 Summary 112 // 4.5 Exercises 113 // 4.6 References 115 // 5 Image pre-processing 116 // 5.1 Pixel brightness transformations 117 // 5.1.1 Position-dependent brightness correction 117 // 5.1.2 Gray-scale transformation 117 // 5.2 Geometric transformations 120 // 5.2.1 Pixel co-ordinate transformations 121 // 5.2.2 Brightness interpolation 123 // 5.3 Local pre-processing 125 // 5.3.1 Image smoothing 125 // 5.3.2 Edge detectors 133 // 5.3.3 Zero-crossings of the second derivative 139 // 5.3.4 Scale in image processing 143 // 5.3.5 Canny edge detection 144 // 5.3.6 Parametric edge models 147 // 5.3.7 Edges in multi-spectral images 148 // 5.3.8 Local pre-processing in the frequency domain ’ 148 // 5.3.9 Line detection by local pre-processing operators 155 // 5.3.10 Detection of corners (interest points) 156 // 5.3.11 Detection of maximally stable extremal regions 160 // 5.4 Image restoration 162 // 5.4.1 Degradations that are easy to restore 163 // 5.4.2 Inverse filtering 163 // 5.4.3 Wiener filtering 164 // 5.5 Summary 165 // 5.6 Exercises 167 // 5.7 References 174 // 6 Segmentation I 178 // 6.1 Thresholding 179 // 6.1.1 Threshold detection methods 181 // 6.1.2 Optimal thresholding 183 // 6.1.3 Multi-spectral thresholding 186 // 6.2 Edge-based segmentation 187 //

6.2.1 Edge image thresholding 188 // 6.2.2 Edge relaxation 190 // 6.2.3 Border tracing 191 // 6.2.4 Border detection as graph searching 196 // 6.2.5 Border detection as dynamic programming 206 // 6.2.6 Hough transforms 210 // 6.2.7 Border detection using border location information 217 // 6.2.8 Region construction from borders 218 // 6.3 Region-based segmentation 220 // 6.3.1 Region merging 221 // 6.3.2 Region splitting 224 // 6.3.3 Splitting and merging 225 // 6.3.4 Watershed segmentation 229 // 6.3.5 Region growing post-processing 232 // 6.4 Matching 232 // 6.4.1 Template matching 233 // 6.4.2 Control strategies of templating 235 // 6.5 Evaluation issues in segmentation 236 // 6.5.1 Supervised evaluation 237 // 6.5.2 Unsupervised evaluation 240 // 6.6 Summary 241 // 6.7 Exercises 245 // 6.8 References 248 // 7 Segmentation II 255 // 7.1 Mean shift segmentation 255 // 7.2 Active contour models—snakes 263 // 7.2.1 Traditional snakes and balloons 264 // 7.2.2 Extensions 267 // 7.2.3 Gradient vector flow snakes 268 // 7.3 Geometric deformable models—level sets and geodesic active contours 273 // 7.4 Fuzzy connectivity 280 // 7.5 Towards 3D graph-based image segmentation 288 // 7.5.1 Simultaneous detection of border pairs 289 // 7.5.2 Suboptimal surface detection 293 // 7.6 Graph cut segmentation 295 // 7.7 Optimal single and multiple surface segmentation— LOGISMOS 303 // 7.8 Summary 317 // 7.9 Exercises 319 // 7.10 References 321 // 8 Shape representation and description 329 // 8.1 Region identification 333 // 8.2 Contour-based shape representation and description 335 // 8.2.1 Chain codes 336 // 8.2.2 Simple geometric border representation 337 // 8.2.3 Fourier transforms of boundaries 341 // 8.2.4 Boundary description using segment sequences 343 // 8.2.5 B-spline representation 346 // 8.2.6 Other contour-based shape description approaches 348 //

8.2.7 Shape invariants 349 // 8.3 Region-based shape representation and description 353 // 8.3.1 Simple scalar region descriptors 353 // 8.3.2 Moments ’ 358 // 8.3.3 Convex hull 360 // 8.3.4 Graph representation based on region skeleton 365 // 8.3.5 Region decomposition 370 // 8.3.6 Region neighborhood graphs 372 // 8.4 Shape classes 373 // 8.5 Summary 373 // 8.6 Exercises 375 // 8.7 References 379 // 9 Object recognition 385 // 9.1 Knowledge representation 386 // 9.2 Statistical pattern recognition 390 // 9.2.1 Classification principles 392 // 9.2.2 Nearest neighbors 393 // 9.2.3 Classifier setting 395 // 9.2.4 Classifier learning 398 // 9.2.5 Support vector machines 400 // 9.2.6 Cluster analysis 406 // 9.3 Neural nets 407 // 9.3.1 Feed-forward networks 409 // 9.3.2 Unsupervised learning 411 // 9.3.3 Hopfield neural nets 412 // 9.4 Syntactic pattern recognition 413 // 9.4.1 Grammars and languages 415 // 9.4.2 Syntactic analysis, syntactic classifier 417 // 9.4.3 Syntactic classifier learning, grammar inference 420 // 9.5 Recognition as graph matching 421 // 9.5.1 Isomorphism of graphs and subgraphs 421 // 9.5.2 Similarity of graphs 425 // 9.6 Optimization techniques in recognition 426 // 9.6.1 Genetic algorithms 427 // 9.6.2 Simulated annealing 430 // 9.7 Fuzzy systems 432 // 9.7.1 Fuzzy sets and fuzzy membership functions 432 // 9.7.2 Fuzzy set operators 434 // 9.7.3 Fuzzy reasoning 435 // 9.7.4 Fuzzy system design and training 438 // 9.8 Boosting in pattern recognition 439 // 9.9 Random forests 442 // 9.9.1 Random forest training 444 // 9.9.2 Random forest decision making 446 // 9.9.3 Random forest extensions 448 // 9.10 Summary 448 // 9.11 Exercises 452 // 9.12 References 459 // 10 Image understanding 464 // 10.1 Image understanding control strategies 466 // 10.1.1 Parallel and serial processing control 466 // 10.1.2 Hierarchical control 466 //

10.1.3 Bottom-up control 467 // 10.1.4 Model-based control 468 // 10.1.5 Combined control 469 // 10.1.6 Non-hierarchical control 472 // 10.2 SIFT: Scale invariant feature transform 474 // 10.3 RANSAC: Fitting via random sample consensus 477 // 10.4 Point distribution models 481 // 10.5 Active appearance models 492 // 10.6 Pattern recognition methods in image understanding 503 // 10.6.1 Classification-based segmentation 503 // 10.6.2 Contextual image classification 505 // 10.6.3 Histograms of oriented gradients—HOG 509 // 10.7 Boosted cascades of classifiers 513 // 10.8 Image understanding using random forests 517 // 10.9 Scene labeling and constraint propagation 524 // 10.9.1 Discrete relaxation 525 // 10.9.2 Probabilistic relaxation 527 // 10.9.3 Searching interpretation trees 530 // 10.10 Semantic image segmentation and understanding 531 // 10.10.1 Semantic region growing 532 // 10.10.2 Genetic image interpretation 534 // 10.11 Hidden Markov models 543 // 10.11.1 Applications 548 // 10.11.2 Coupled HMMs 549 // 10.11.3 Bayesian belief networks 551 // 10.12 Markov random fields 553 // 10.12.1 Applications to images and vision 555 // 10.13 Gaussian mixture models and expectation-maximization 556 // 10.14 Summary 564 // 10.15 Exercises 568 // 10.16 References 572 // 11 3D geometry, correspondence, 3D from intensities 582 // 11.1 3D vision tasks 583 // 11.1.1 Marr’s theory 585 // 11.1.2 Other vision paradigms: Active and purposive vision 587 // 11.2 Basics of projective geometry 589 // 11.2.1 Points and hyperplanes in projective space 590 // 11.2.2 Homography 592 // 11.2.3 Estimating homography from point correspondences 594 // 11.3 A single perspective camera 598 // 11.3.1 Camera model 598 // 11.3.2 Projection and back-projection in homogeneous coordinates 601 // 11.3.3 Camera calibration from a known scene 602 //

11.4 Scene reconstruction from multiple views 602 // 11.4.1 Triangulation 603 // 11.4.2 Projective reconstruction 604 // 11.4.3 Matching constraints 605 // 11.4.4 Bundle adjustment 607 // 11.4.5 Upgrading the projective reconstruction, self-calibration 608 // 11.5 Two cameras, stereopsis 609 // 11.5.1 Epipolar geometry; fundamental matrix 610 // 11.5.2 Relative motion of the camera; essential matrix 612 // 11.5.3 Decomposing the fundamental matrix to camera matrices 613 // 11.5.4 Estimating the fundamental matrix from point correspondences 614 // 11.5.5 Rectified configuration of two cameras 615 // 11.5.6 Computing rectification 617 // 11.6 Three cameras and trifocal tensor 619 // 11.6.1 Stereo correspondence algorithms 621 // 11.6.2 Active acquisition of range images 627 // 11.7 3D information from radiometric measurements 630 // 11.7.1 Shape from shading 631 // 11.7.2 Photometric stereo 635 // 11.8 Summary 636 // 11.9 Exercises 637 // 11.10 References . 639 // 12 Use of 3D vision 644 // 12.1 Shape from X 644 // 12.1.1 Shape from motion 644 // 12.1.2 Shape from texture 651 // 12.1.3 Other shape from X techniques 652 // 12.2 Full 3D objects 655 // 12.2.1 3D objects, models, and related issues 655 // 12.2.2 Line labeling 656 // 12.2.3 Volumetric representation, direct measurements 658 // 12.2.4 Volumetric modeling strategies 660 // 12.2.5 Surface modeling strategies 662 // 12.2.6 Registering surface patches and their fusion to get a full 3D // model 663 // 12.3 2D view-based representations of a 3D scene 670 // 12.3.1 Viewing space 670 // 12.3.2 Multi-view representations and aspect graphs 670 // 12.4 3D reconstruction from an unorganized set of 2D views, and Structure // from Motion 671 // 12.5 Reconstructing scene geometry 674 // 12.6 Summary 677 // 12.7 Exercises 678 // 12.8 References 680 // 13 Mathematical morphology 684 //

13.1 Basic morphological concepts 684 // 13.2 Four morphological principles 686 // 13.3 Binary dilation and erosion 687 // 13.3.1 Dilation 688 // 13.3.2 Erosion 689 // 13.3.3 Hit-or-miss transformation 692 // 13.3.4 Opening and closing 692 // 13.4 Gray-scale dilation and erosion 694 // 13.4.1 Top surface, umbra, and gray-scale dilation and erosion 694 // 13.4.2 Umbra homeomorphism theorem, properties of erosion and // dilation, opening and closing 697 // 13.4.3 Top hat transformation 698 // 13.5 Skeletons and object marking 699 // 13.5.1 Homotopic transformations 699 // 13.5.2 Skeleton, mediai axis, maximal ball 699 // 13.5.3 Thinning, thickening, and homotopic skeleton 701 // 13.5.4 Quench function, ultimate erosion 704 // 13.5.5 Ultimate erosion and distance functions 706 // 13.5.6 Geodesic transformations 707 // 13.5.7 Morphological reconstruction 709 // 13.6 Granulometry 711 // 13.7 Morphological segmentation and watersheds 713 // 13.7.1 Particle segmentation, marking, and watersheds 713 // 13.7.2 Binary morphological segmentation 714 // 13.7.3 Gray-scale segmentation, watersheds 716 // 13.8 Summary 717 // 13.9 Exercises 718 // 13.10 References 720 // 14 Image data compression 722 // 14.1 Image data properties 723 // 14.2 Discrete image transforms in image data compression 724 // 14.3 Predictive compression methods 727 // 14.4 Vector quantization 730 // 14.5 Hierarchical and progressive compression methods 730 // 14.6 Comparison of compression methods 732 // 14.7 Other techniques 733 // 14.8 Coding 733 // 14.9 JPEG and MPEG image compression 734 // 14.9.1 JPEG—still image compression 734 // 14.9.2 JPEG-2000 compression 736 // 14.9.3 MPEG—full-motion video compression 738 // 14.10 Summary 740 // 14.11 Exercises 742 // 14.12 References 744 // 15 Texture 747 // 15.1 Statistical texture description 750 // 15.1.1 Methods based on spatial frequencies 750 //

15.1.2 Co-occurrence matrices 752 // 15.1.3 Edge frequency 754 // 15.1.4 Primitive length (run length) 755 // 15.1.5 Laws’ texture energy measures 757 // 15.1.6 Local binary patterns—LBPs 757 // 15.1.7 Fractal texture description 762 // 15.1.8 Multiscale texture description—wavelet domain approaches 764 // 15.1.9 Other statistical methods of texture description 768 // 15.2 Syntactic texture description methods 769 // 15.2.1 Shape chain grammars 770 // 15.2.2 Graph grammars 772 // 15.2.3 Primitive grouping in hierarchical textures 773 // 15.3 Hybrid texture description methods 775 // 15.4 Texture recognition method applications 776 // 15.5 Summary 777 // 15.6 Exercises 779 // 15.7 References 782 // 16 Motion analysis 787 // 16.1 Differential motion analysis methods 790 // 16.2 Optical flow 794 // 16.2.1 Optical flow computation 794 // 16.2.2 Global and local optical flow estimation 797 // 16.2.3 Combined local-global optical flow estimation 800 // 16.2.4 Optical flow in motion analysis 801 // 16.3 Analysis based on correspondence of interest points 804 // 16.3.1 Detection of interest points 805 // 16.3.2 Lucas-Kanade point tracking 805 // 16.3.3 Correspondence of interest points 807 // 16.4 Detection of specific motion patterns 810 // 16.5 Video tracking 814 // 16.5.1 Background modeling 815 // 16.5.2 Kernel-based tracking 820 // 16.5.3 Object path analysis 826 // 16.6 Motion models to aid tracking 831 // 16.6.1 Kalman filters 831 // 16.6.2 Particle filters 837 // 16.6.3 Semi-supervised tracking—TLD 840 // 16.7 Summary 843 // 16.8 Exercises 846 // 16.9 References 848 // Index 853