Main description:

Medical imaging is an important topic and plays a key role in robust diagnosis and patient care. It has experienced an explosive growth over the last few years due to imaging modalities such as X-rays, computed tomography (CT), magnetic resonance (MR) imaging, and ultrasound. This book focuses primarily on model-based segmentation techniques, which are applied to cardiac, brain, breast and microscopic cancer cell imaging. It includes contributions from authors working in industry and academia, and presents new material.

Contents:

1. Principles of Image Generation.- 1.1 Introduction.- 1.2 Ultrasound Image Generation.- 1.2.1 The Principle of Pulse-Echo Ultrasound Imaging.- 1.2.2 B-Scan Quality and the Ultimate Limits.- 1.2.3 Propagation-Related Artifacts and Resolution Limits.- 1.2.3 Attenuation-Related Artifacts.- 1.3 X-Ray Cardiac Image Generation.- 1.3.1 LV Data Acquisition System Using X-Rays.- 1.3.2 Drawbacks of Cardiac Catheterization.- 1.4 Magnetic Resonance Image Generation.- 1.4.1 Physical Principles of Nuclear Magnetic Resonance.- 1.4.2 Basics of Magnetic Resonance Imaging.- 1.4.3 Gradient-Echo (GRE).- 1.4.4 The Latest Techniques for MR Image Generation.- 1.4.5 3-D Turbo FLASH (MP-RAGE) Technique.- 1.4.6 Non-Rectilinear k-Space Trajectory: Spiral.- 1.4.7 Fat Suppression.- 1.4.8 High Speed MRI: Perfusion-Weighted.- 1.4.9 Time of Flight (TOF) MR Angiography.- 1.4.10 Fast Spectroscopic Imaging.- 1.4.11 Recent MR Imaging Techniques.- 1.5 Computer Tomography Image Generation.- 1.5.1 Fourier Reconstruction Method.- 1.6 Positron-Emission Tomography Image Generation.- 1.6.1 Underlying Principles of.- 1.6.2 Usage of PET in Diagnosis.- 1.6.3 Fourier Slice Theorem.- 1.6.4 The Reconstruction Algorithm in PET.- 1.6.5 Image Reconstruction Using Filtered Back-Projection.- 1.7 Comparison of Imaging Modalities: A Summary.- 1.7.1 Acknowledgements.- 2. Segmentation in Echocardiographic Images.- 2.1 Introduction.- 2.2 Heart Physiology and Anatomy.- 2.2.1 Cardiac Function.- 2.2.2 Standard LV Views in 2-DEs.- 2.2.3 LV Function Assessment Using 2-DEs.- 2.3 Review of LV Boundary Extraction Techniques Applied to Echocardiographic Data.- 2.3.1 Acoustic Quantification Techniques.- 2.3.2 Image-Based Techniques.- 2.3.3 2-DE Image Processing Techniques.- 2.4Automatic Fuzzy Reasoning-Based Left Ventricular Center Point Extraction.- 2.4.1 LVCP Extraction System Overview.- 2.4.2 Stage 1: Pre-Processing.- 2.4.3 Stage 2: LVCP Features Fuzzification.- 2.4.4 Template Matching.- 2.4.5 Experimental Results.- 2.4.6 Conclusion.- 2.5 A New Edge Detection in the Wavelet Transform Domain.- 2.5.1 Multiscale Edge Detection and the Wavelet Transform.- 2.5.2 Edge Detection Based on the Global Maximum of Wavelet Transform (GMWT).- 2.5.3 GMWT Performance Analysis and Comparison.- 2.6 LV Segmentation System.- 2.6.1 Overall Reference.- 2.6.2 3D Non-Uniform Radial Intensity Sampling.- 2.6.3 LV Boundary Edge Detection on 3D Radial Intensity Matrix.- 2.6.4 Post-Processing of the Edges and Closed LVE Approximation.- 2.6.5 Automatic LV Volume Assessment.- 2.7 Conclusions.- 2.8 Acknowledgments.- 3. Cardiac Boundary Segmentation.- 3.1 Introduction.- 3.2 Cardiac Anatomy and Data Acquisitions for MR, CT, Ul-trasound and X-Rays.- 3.2.1 Cardiac Anatomy.- 3.2.2 Cardiac MR, CT, Ultrasound and X-Ray Acquisitions.- 3.3 Low- and Medium-Level LV Segmentation Techniques.- 3.3.1 Smoothing Image Data.- 3.3.2 Manual and Semi-Automatic LV Thresholding.- 3.3.3 LV Dynamic Thresholding.- 3.3.4 Edge-Based Techniques.- 3.3.5 Mathematical Morphology-Based Techniques.- 3.3.6 Drawbacks of Low-Level LV Segmentation Techniques.- 3.4 Model-Based Pattern Recognition Methods for LV Modeling.- 3.4.1 LV Active Contour Models in the Spatial and Temporal Domains.- 3.4.2 Model-Based Pattern Recognition Learning Methods.- 3.4.3 Polyline Distance Measure and Performance Terms.- 3.4.4 Data Analysis Using IdCM, InCM and the Greedy Method.- 3.5 Left Ventricle Apex Modeling: A Model-Based Approach.- 3.5.1 Longitudinal Axis and Apex Modeling.- 3.5.2 Ruled Surface Model.- 3.5.3 Ruled Surface sr and its Coefficients.- 3.5.4 Estimation of Robust Coefficients and Coordinates of the Ruled Surface.- 3.5.5 Experiment Design.- 3.5.6 Analytical Error Measure, AQin for Inlier Data.- 3.5.7 Experiments, Results and Discussions.- 3.5.8 Conclusions on LV Apex Modeling.- 3.6 Integration of Low-Level Features in LV Model-Based Cardiac Imaging: Fusion of Two Computer Vision Systems.- 3.7 General Purpose LV Validation Technique.- 3.8 LV Convex Hulling: Quadratic Training-Based Point Modeling.- 3.8.1 Quadratic Vs. Linear Optimization for Convex Hulling.- 3.9 LV Eigen Shape Modeling.- 3.9.1 Procrustes Superposition.- 3.9.2 Dimensionality Reduction Using Constraints for Joint.- 3.10 LV Neural Network Models.- 3.11 Comparative Study and Summary of the Characteristics of Model-Based Techniques.- 3.11.1 Characteristics of Model-Based LV Imaging.- 3.12 LV Quantification: Wall Motion and Tracking.- 3.12.1 LV Wall Motion Measurements.- 3.12.2 LV Volume Measurements.- 3.12.3 LV Wall Motion Tracking.- 3.13 Conclusions.- 3.13.1 Cardiac Hardware.- 3.13.2 Cardiac Software.- 3.13.3 Summary.- 3.13.4 Acknowledgments.- 4. Brain Segmentation Techniques.- 4.1 Introduction.- 4.1.1 Human Brain Anatomy and the MRI System.- 4.1.2 Applications of Brain Segmentation.- 4.2 Brain Scanning and its Clinical Significance.- 4.3 Region-Based 2-D and 3-D Cortical Segmentation Techniques.- 4.3.1 Atlas-Based and Threshold-Based Techniques.- 4.3.2 Cortical Segmentation Using Probability-Based Techniques.- 4.3.3 Clustering-Based Cortical Segmentation Techniques.- 4.3.4 Mathematical Morphology-Based Cortical Segmentation Techniques.- 4.3.5 Prior Knowledge-Based Techniques.- 4.3.6 Texture-Based Techniques.- 4.3.7 Neural Network-Based Techniques.- 4.3.8 Regional Hyperstack: Fusion of Edge-Diffusion with Region-Linking.- 4.3.9 Fusion of Probability-Based with Edge Detectors, Connectivity and Region-Growing.- 4.3.10 Summary of Region-Based Techniques: Pros and Cons.- 4.4 Boundary/Surface-Based 2-D and 3-D Cortical Segmentation Techniques: Edge, Reconstruction, Parametric and Geometric Snakes/Surfaces.- 4.4.1 Edge-Based Cortical-Boundary Estimation Techniques.- 4.4.2 3-D Cortical Reconstruction From 2-D Serial Cross-Sections (Bourke/Victoria).- 4.4.3 2-D and 3-D Parametric Deformable Models for Cortical Boundary Estimation: Snakes, Fitting, Constrained, Ribbon, T-Surface, Connectedness.- 4.4.4 2-D and 3-D Geometric Deformable Models.- 4.4.5 A Note on Isosurface Extraction (Lorensen/GE).- 4.4.6 Summary of Boundary/Surface-Based Techniques: Pros and Cons.- 4.5 Fusion of Boundary/Surface with Region-Based 2-D and 3-D Cortical Segmentation Techniques.- 4.5.1 2-D/3-D Regional Parametric Boundary: Fusion of Boundary with Classification (Kapur/MIT).- 4.5.2 Regional Parametric Surfaces: Fusion of Surface with Clustering (Xu/JHU).- 4.5.3 2-D Regional Geometric Boundary: Fusion of Boundary with Clustering for Cortical Boundary Estimation (Suri/Marconi).- 4.5.34 3-D Regional Geometric Surfaces: Fusion of Geometric Surface with Probability-Based Voxel Classification (Zeng/Yale).- 4.5.5 2-D/3-D Regional Geometric Surface: Fusion of Geometric Boundary/Surface with Global Shape Information (Leventon/MIT).- 4.5.6 2-D/3-D Regional Geometric Surface: Fusion of Boundary/Surface with Bayesian-Based Pixel Classification (Barillot/IRISA).- 4.5.7 Similarities/Differences Between Different Cortical Segmentation Techniques.- 4.6 3-D Visualization Using Volume Rendering and Texture Mapping.- 4.6.1 Volume Rendering Algorithm for Brain Segmentation.- 4.6.2 Texture Mapping Algorithm for Segmented Brain Visualization.- 4.7 A Note on fMRI: Algorithmic Approach for Establishing the Relationship Between Cognitive Functions and Brain Cortical Anatomy.- 4.7.1 Superiority of fMRI over PET/SPECT Imaging.- 4.7.2 Applications of fMRI.- 4.7.3 Algorithm for Superimposition of Functional and Anatomical Cortex.- 4.7.4 A Short Note on fMRI Time Course Data Analysis.- 4.7.5 Measure of Cortex Geometry.- 4.8 Discussions: Advantages, Validation and New Challenges i 2-D.- 4.8.1 Advantages of Regional Geometric Boundary/Surfaces.- 4.8.2 Validation of 2-D and 3-D Cortical Segmentation Algorithms.- 4.8.3 Challenges in 2-D and 3-D Cortical Segmentation Algorithms.- 4.8.4 Challenges in fMRI.- 4.9 Conclusions and the Future.- 4.9.1 Acknowledgements.- 5. Segmentation for Multiple Sclerosis Lesion.- 5.1 Introduction.- 5.2 Segmentation Techniques.- 5.2.1 Multi-Spectral Techniques.- 5.2.2 Feature Space Classification.- 5.2.3 Supervised Segmentation.- 5.2.4 Unsupervised Segmentation.- 5.2.5 Automatic Segmentation.- 5.3 AFFIRMATIVE Images.- 5.4 Image Pre-Processing.- 5.4.1 RF Inhomogeneity Correction.- 5.4.2 Image Stripping.- 5.4.3 Three Dimensional MR Image Registration.- 5.4.4 Segmentation.- 5.4.5 Flow Correction.- 5.4.6 Evaluation and Validation.- 5.5 Quantification of Enhancing Multiple Sclerosis Lesions.- 5.6 Quadruple Contrast Imaging.- 5.7 Discussion.- 5.7.1 Acknowledgements.- 6. Finite Mixture Models.- 6.1 Introduction.- 6.2 Pixel Labeling Using the Classical Mixture Model.- 6.3 Pixel Labeling Using the Spatially Variant Mixture Model.- 6.4 Comparison of CMM and SVMM for Pixel Labeling.- 6.5 Bayesian Pixel Labeling Using the SVMM.- 6.6 Segmentation Results.- 6.6.1 Computer Simulations.- 6.6.2 Application to Magnetic Resonance Images.- 6.7 Practical Aspects.- 6.8 Summary.- 6.9 Acknowledgements.- 7. MR Spectroscopy.- 7.1 Introduction.- 7.2 A Short History of Neurospectroscopic Imaging and Segmentation in Alzheimer's Disease and Multiple Sclerosis.- 7.2.1 Alzheimer's Disease.- 7.2.2 Multiple Sclerosis.- 7.3 Data Acquisition and Image Segmentation.- 7.3.1 Image Pre-Processing for Segmentation.- 7.3.2 Image Post-Processing for Segmentation.- 7.4 Proton Magnetic Resonance Spectroscopic Imaging and Segmentation in Multiple Sclerosis.- 7.4.1 Automatic MRSI Segmentation and Image Processing Algorithm.- 7.4.2 Relative Metabolite Concentrations and Contribution of Gray Matter and White Matter in the Normal Human Brain.- 7.4.3 MRSI and Gadolinium-Enhanced (Gd).- 7.4.4 Lesion Load and Metabolite Concentrations by Segmentation and MRSI.- 7.4.5 MR Spectroscopic Imaging and Localization for Segmentation.- 7.4.6 Lesion Segmentation and Quantification.- 7.4.7 Magnetic Resonance Spectroscopic Imaging and Segmentation Data Processing.- 7.4.8 Statistical Analysis.- 7.5 Proton Magnetic Resonance Spectroscopic Imaging and Segmentation of Alzheimer's Disease.- 7.5.1 MRSI Data Acquisition Methods.- 7.5.2 H-1 MR Spectra Analysis.- 7.6 Applications of Magnetic Resonance Spectroscopic Imaging and Segmentation.- 7.6.1 Multiple Sclerosis Lesion Metabolite Characteristics and Serial Changes.- 7.6.2 zheimer's Disease Plaque Metabolite Characteristics.- 7.7 Discussion.- 7.8 Conclusion.- 7.8.1 Acknowledgements.- 8. Fast WM/GM Boundary Estimation.- 8.1 Introduction.- 8.2 Derivation of the Regional Geometric Active Contour Model from the Classical Parametric Deformable Model.- 8.3 Numerical Implementation of the Three Speed Functions in the Level Set Framework for Geometric Snake Propagation.- 8.3.1 Regional Speed Term Expressed in Terms of the Level Set Function (o).- 8.3.2 Gradient Speed Term Expressed in Terms of the Level Set Function (o).- 8.3.3 Curvature Speed Term Expressed in Terms of the Level Set Function (o).- 8.4 Fast Brain Segmentation System Based on Regional Level Sets.- 8.4.1 Overall System and Its Components.- 8.4.2 Fuzzy Membership Computation/Pixel Classification.- 8.4.3 Eikonal Equation and its Mathematical Solution.- 8.4.4 Fast Marching Method for Solving the Eikonal Equation.- 8.4.5 A Note on the Heap Sorting Algorithm.- 8.4.6 Segmentation Engine: Running the Level Set Method in the Narrow Band.- 8.5 MR Segmentation Results on Synthetic and Real Data.- 8.5.1 Input Data Set and Input Level Set Parameters.- 8.5.2 Results: Synthetic and Real.- 8.5.3 Numerical Stability, Signed Distance Transformation Computation, Sensitivity of Parameters and Speed Issues.- 8.6 Advantages of the Regional Level Set Technique.- 8.7 Discussions: Comparison with Previous Techniques.- 8.8 Conclusions and Further Directions.- 8.8.1 Acknowledgements.- 9. Digital Mammography Segmentation.- 9.1 Introduction.- 9.2 Image Segmentation in Mammography.- 9.3 Anatomy of the Breast.- 9.4 Image Acquisition and Formats.- 9.4.1 Digitization of X-Ray Mammograms.- 9.4.2 Image Formats.- 9.4.3 Image Quantization and Tree-Pyramids.- 9.5 Mammogram Enhancement Methods.- 9.6 Quantifying Mammogram Enhancement.- 9.7 Segmentation of Breast Profile.- 9.8 Segmentation of Microcalcifications.- 9.9 Segmentation of Masses.- 9.9.1 Global Methods.- 9.9.2 Edge-Based Methods.- 9.9.3 Region-Based Segmentation.- 9.9.4 ROI Detection Techniques Using a Single Breast.- 9.9.5 ROI Detection Techniques Using Breast Symmetry.- 9.9.6 Detection of Spicules.- 9.9.7 Breast Alignment for Segmentation.- 9.10 Measures of Segmentation and Abnormality Detection.- 9.11 Feature Extraction From Segmented Regions.- 9.11.1 Morphological Features.- 9.11.2 Texture Features.- 9.11.3 Other Features.- 9.12 Public Domain Databases in Mammography.- 9.12.1 The Digital Database for Screening Mammography (DDSM).- 9.12.2 LLNL/UCSF Database.- 9.12.3 Washington University Digital Mammography Database.- 9.12.4 The Mammographic Image Analysis Society (MIAS) Database.- 9.13 Classification and Measures of Performance.- 9.13.1 Classification Techniques.- 9.13.2 The Receiver Operating Characteristic Curve.- 9.14 Conclusions.- 9.15 Acknowledgements.- 10. Cell Image Segmentation for Diagnostic Pathology.- 10.1 Introduction.- 10.2 Segmentation.- 10.2.1 Feature Space Analysis.- 10.2.2 Mean Shift Procedure.- 10.2.3 Cell Segmentation.- 10.2.4 Segmentation Examples.- 10.3 Decision Support System for Pathology.- 10.3.1 Problem Domain.- 10.3.2 System Overview.- 10.3.3 Current Database.- 10.3.4 Analysis of Visual Attributes.- 10.3.5 Overall Dissimilarity Metric.- 10.3.6 Performance Evaluation and Comparisons.- 10.4 Conclusion.- 11. The Future in Segmentation.- 11.1 Future Research in Medical Image Segmentation.- 11.1.1 The Future of MR Image Generation and Physical Principles.- 11.1.2 The Future of Cardiac Imaging.- 11.2.3 The Future of Neurological Segmentation.- 11.2.4 The Future in Digital Mammography.- 11.2.5 The Future of Pathology Image Segmentation.