- •Biological and Medical Physics, Biomedical Engineering
- •Medical Image Processing
- •Preface
- •Contents
- •Contributors
- •1.1 Medical Image Processing
- •1.2 Techniques
- •1.3 Applications
- •1.4 The Contribution of This Book
- •References
- •2.1 Introduction
- •2.2 MATLAB and DIPimage
- •2.2.1 The Basics
- •2.2.2 Interactive Examination of an Image
- •2.2.3 Filtering and Measuring
- •2.2.4 Scripting
- •2.3 Cervical Cancer and the Pap Smear
- •2.4 An Interactive, Partial History of Automated Cervical Cytology
- •2.5 The Future of Automated Cytology
- •2.6 Conclusions
- •References
- •3.1 The Need for Seed-Driven Segmentation
- •3.1.1 Image Analysis and Computer Vision
- •3.1.2 Objects Are Semantically Consistent
- •3.1.3 A Separation of Powers
- •3.1.4 Desirable Properties of Seeded Segmentation Methods
- •3.2 A Review of Segmentation Techniques
- •3.2.1 Pixel Selection
- •3.2.2 Contour Tracking
- •3.2.3 Statistical Methods
- •3.2.4 Continuous Optimization Methods
- •3.2.4.1 Active Contours
- •3.2.4.2 Level Sets
- •3.2.4.3 Geodesic Active Contours
- •3.2.5 Graph-Based Methods
- •3.2.5.1 Graph Cuts
- •3.2.5.2 Random Walkers
- •3.2.5.3 Watershed
- •3.2.6 Generic Models for Segmentation
- •3.2.6.1 Continuous Models
- •3.2.6.2 Hierarchical Models
- •3.2.6.3 Combinations
- •3.3 A Unifying Framework for Discrete Seeded Segmentation
- •3.3.1 Discrete Optimization
- •3.3.2 A Unifying Framework
- •3.3.3 Power Watershed
- •3.4 Globally Optimum Continuous Segmentation Methods
- •3.4.1 Dealing with Noise and Artifacts
- •3.4.2 Globally Optimal Geodesic Active Contour
- •3.4.3 Maximal Continuous Flows and Total Variation
- •3.5 Comparison and Discussion
- •3.6 Conclusion and Future Work
- •References
- •4.1 Introduction
- •4.2 Deformable Models
- •4.2.1 Point-Based Snake
- •4.2.1.1 User Constraint Energy
- •4.2.1.2 Snake Optimization Method
- •4.2.2 Parametric Deformable Models
- •4.2.3 Geometric Deformable Models (Active Contours)
- •4.2.3.1 Curve Evolution
- •4.2.3.2 Level Set Concept
- •4.2.3.3 Geodesic Active Contour
- •4.2.3.4 Chan–Vese Deformable Model
- •4.3 Comparison of Deformable Models
- •4.4 Applications
- •4.4.1 Bone Surface Extraction from Ultrasound
- •4.4.2 Spinal Cord Segmentation
- •4.4.2.1 Spinal Cord Measurements
- •4.4.2.2 Segmentation Using Geodesic Active Contour
- •4.5 Conclusion
- •References
- •5.1 Introduction
- •5.2 Imaging Body Fat
- •5.3 Image Artifacts and Their Impact on Segmentation
- •5.3.1 Partial Volume Effect
- •5.3.2 Intensity Inhomogeneities
- •5.4 Overview of Segmentation Techniques Used to Isolate Fat
- •5.4.1 Thresholding
- •5.4.2 Selecting the Optimum Threshold
- •5.4.3 Gaussian Mixture Model
- •5.4.4 Region Growing
- •5.4.5 Adaptive Thresholding
- •5.4.6 Segmentation Using Overlapping Mosaics
- •5.6 Conclusions
- •References
- •6.1 Introduction
- •6.2 Clinical Context
- •6.3 Vessel Segmentation
- •6.3.1 Survey of Vessel Segmentation Methods
- •6.3.1.1 General Overview
- •6.3.1.2 Region-Growing Methods
- •6.3.1.3 Differential Analysis
- •6.3.1.4 Model-Based Filtering
- •6.3.1.5 Deformable Models
- •6.3.1.6 Statistical Approaches
- •6.3.1.7 Path Finding
- •6.3.1.8 Tracking Methods
- •6.3.1.9 Mathematical Morphology Methods
- •6.3.1.10 Hybrid Methods
- •6.4 Vessel Modeling
- •6.4.1 Motivation
- •6.4.1.1 Context
- •6.4.1.2 Usefulness
- •6.4.2 Deterministic Atlases
- •6.4.2.1 Pioneering Works
- •6.4.2.2 Graph-Based and Geometric Atlases
- •6.4.3 Statistical Atlases
- •6.4.3.1 Anatomical Variability Handling
- •6.4.3.2 Recent Works
- •References
- •7.1 Introduction
- •7.2 Linear Structure Detection Methods
- •7.3.1 CCM for Imaging Diabetic Peripheral Neuropathy
- •7.3.2 CCM Image Characteristics and Noise Artifacts
- •7.4.1 Foreground and Background Adaptive Models
- •7.4.2 Local Orientation and Parameter Estimation
- •7.4.3 Separation of Nerve Fiber and Background Responses
- •7.4.4 Postprocessing the Enhanced-Contrast Image
- •7.5 Quantitative Analysis and Evaluation of Linear Structure Detection Methods
- •7.5.1 Methodology of Evaluation
- •7.5.2 Database and Experiment Setup
- •7.5.3 Nerve Fiber Detection Comparison Results
- •7.5.4 Evaluation of Clinical Utility
- •7.6 Conclusion
- •References
- •8.1 Introduction
- •8.2 Methods
- •8.2.1 Linear Feature Detection by MDNMS
- •8.2.2 Check Intensities Within 1D Window
- •8.2.3 Finding Features Next to Each Other
- •8.2.4 Gap Linking for Linear Features
- •8.2.5 Quantifying Branching Structures
- •8.3 Linear Feature Detection on GPUs
- •8.3.1 Overview of GPUs and Execution Models
- •8.3.2 Linear Feature Detection Performance Analysis
- •8.3.3 Parallel MDNMS on GPUs
- •8.3.5 Results for GPU Linear Feature Detection
- •8.4.1 Architecture and Implementation
- •8.4.2 HCA-Vision Features
- •8.4.3 Linear Feature Detection and Analysis Results
- •8.5 Selected Applications
- •8.5.1 Neurite Tracing for Drug Discovery and Functional Genomics
- •8.5.2 Using Linear Features to Quantify Astrocyte Morphology
- •8.5.3 Separating Adjacent Bacteria Under Phase Contrast Microscopy
- •8.6 Perspectives and Conclusions
- •References
- •9.1 Introduction
- •9.2 Bone Imaging Modalities
- •9.2.1 X-Ray Projection Imaging
- •9.2.2 Computed Tomography
- •9.2.3 Magnetic Resonance Imaging
- •9.2.4 Ultrasound Imaging
- •9.3 Quantifying the Microarchitecture of Trabecular Bone
- •9.3.1 Bone Morphometric Quantities
- •9.3.2 Texture Analysis
- •9.3.3 Frequency-Domain Methods
- •9.3.4 Use of Fractal Dimension Estimators for Texture Analysis
- •9.3.4.1 Frequency-Domain Estimation of the Fractal Dimension
- •9.3.4.2 Lacunarity
- •9.3.4.3 Lacunarity Parameters
- •9.3.5 Computer Modeling of Biomechanical Properties
- •9.4 Trends in Imaging of Bone
- •References
- •10.1 Introduction
- •10.1.1 Adolescent Idiopathic Scoliosis
- •10.2 Imaging Modalities Used for Spinal Deformity Assessment
- •10.2.1 Current Clinical Practice: The Cobb Angle
- •10.2.2 An Alternative: The Ferguson Angle
- •10.3 Image Processing Methods
- •10.3.1 Previous Studies
- •10.3.2 Discrete and Continuum Functions for Spinal Curvature
- •10.3.3 Tortuosity
- •10.4 Assessment of Image Processing Methods
- •10.4.1 Patient Dataset and Image Processing
- •10.4.2 Results and Discussion
- •10.5 Summary
- •References
- •11.1 Introduction
- •11.2 Retinal Imaging
- •11.2.1 Features of a Retinal Image
- •11.2.2 The Reason for Automated Retinal Analysis
- •11.2.3 Acquisition of Retinal Images
- •11.3 Preprocessing of Retinal Images
- •11.4 Lesion Based Detection
- •11.4.1 Matched Filtering for Blood Vessel Segmentation
- •11.4.2 Morphological Operators in Retinal Imaging
- •11.5 Global Analysis of Retinal Vessel Patterns
- •11.6 Conclusion
- •References
- •12.1 Introduction
- •12.1.1 The Progression of Diabetic Retinopathy
- •12.2 Automated Detection of Diabetic Retinopathy
- •12.2.1 Automated Detection of Microaneurysms
- •12.3 Image Databases
- •12.4 Tortuosity
- •12.4.1 Tortuosity Metrics
- •12.5 Tracing Retinal Vessels
- •12.5.1 NeuronJ
- •12.5.2 Other Software Packages
- •12.6 Experimental Results and Discussion
- •12.7 Summary and Future Work
- •References
- •13.1 Introduction
- •13.2 Volumetric Image Visualization Methods
- •13.2.1 Multiplanar Reformation (2D slicing)
- •13.2.2 Surface-Based Rendering
- •13.2.3 Volumetric Rendering
- •13.3 Volume Rendering Principles
- •13.3.1 Optical Models
- •13.3.2 Color and Opacity Mapping
- •13.3.2.2 Transfer Function
- •13.3.3 Composition
- •13.3.4 Volume Illumination and Illustration
- •13.4 Software-Based Raycasting
- •13.4.1 Applications and Improvements
- •13.5 Splatting Algorithms
- •13.5.1 Performance Analysis
- •13.5.2 Applications and Improvements
- •13.6 Shell Rendering
- •13.6.1 Application and Improvements
- •13.7 Texture Mapping
- •13.7.1 Performance Analysis
- •13.7.2 Applications
- •13.7.3 Improvements
- •13.7.3.1 Shading Inclusion
- •13.7.3.2 Empty Space Skipping
- •13.8 Discussion and Outlook
- •References
- •14.1 Introduction
- •14.1.1 Magnetic Resonance Imaging
- •14.1.2 Compressed Sensing
- •14.1.3 The Role of Prior Knowledge
- •14.2 Sparsity in MRI Images
- •14.2.1 Characteristics of MR Images (Prior Knowledge)
- •14.2.2 Choice of Transform
- •14.2.3 Use of Data Ordering
- •14.3 Theory of Compressed Sensing
- •14.3.1 Data Acquisition
- •14.3.2 Signal Recovery
- •14.4 Progress in Sparse Sampling for MRI
- •14.4.1 Review of Results from the Literature
- •14.4.2 Results from Our Work
- •14.4.2.1 PECS
- •14.4.2.2 SENSECS
- •14.4.2.3 PECS Applied to CE-MRA
- •14.5 Prospects for Future Developments
- •References
- •15.1 Introduction
- •15.2 Acquisition of DT Images
- •15.2.1 Fundamentals of DTI
- •15.2.2 The Pulsed Field Gradient Spin Echo (PFGSE) Method
- •15.2.3 Diffusion Imaging Sequences
- •15.2.4 Example: Anisotropic Diffusion of Water in the Eye Lens
- •15.2.5 Data Acquisition
- •15.3 Digital Processing of DT Images
- •15.3.2 Diagonalization of the DT
- •15.3.3 Gradient Calibration Factors
- •15.3.4 Sorting Bias
- •15.3.5 Fractional Anisotropy
- •15.3.6 Other Anisotropy Metrics
- •15.4 Applications of DTI to Articular Cartilage
- •15.4.1 Bovine AC
- •15.4.2 Human AC
- •References
- •Index
5 Fat Segmentation in Magnetic Resonance Images |
105 |
Fig. 5.10 (a) A T1w GE image and the seed points used for region growing and (b) is the resultant segmentation following region growing, visceral fat is labeled with red, other fat is labeled blue and the liver purple. Each labeled group is quantified separately
to the lack of clearly defined edges. Despite this, the data presented by Brennan et al. were well segmented. Brennan’s method did not classify and label body fat. Further steps are required to develop classification algorithm.
5.4.5 Adaptive Thresholding
Due to intensity inhomogeneities in MR images a single threshold may not be sufficient for the entire image. Segmentation using adaptive thresholding can
106 |
D.P. Costello and P.A. Kenny |
Fig. 5.11 (a) Whole body T1-weighted GE image affected by intensity inhomogeneities; (b) Result of global segmentation using the GMM on (a); (c) Sub-images used for adaptive thresholding; and (d) is the result of adaptive thresholding
compensate for intensity inhomogeneities [39]. Adaptive segmentation can be achieved by dividing an image into a number sub-images as shown in Fig. 5.11c [50]. Each sub-image is then segmented using one of the segmentation algorithms discussed in Sect. 5.4.2.
Two factors must be considered when selecting the size of the sub-images:
(1)They must be small enough so the impact of the intensity inhomogeneity is minimal across each of their areas
(2)They must contain enough voxels to maintain a workable SNR
Figure 5.11a is an example of an image that is affected by intensity inhomogeneities and (b) is the result of global segmentation using the GMM algorithm. Using adaptive segmentation a significant improvement can be seen in Fig. 5.11d. If the sub-images cannot be made small enough to reduce the impact of the intensity inhomogeneities, a technique which uses overlapping mosaics may be used [35], this is discussed in Sect. 5.4.6.
Local adaptive thresholding (using a sliding window) can be an effective segmentation algorithm in the presence of inhomogeneities [51]. This technique
5 Fat Segmentation in Magnetic Resonance Images |
107 |
thresholds individual voxels using the mean or median value of their surrounding n × n neighborhood. MR images acquired for fat analysis can contain large monotone regions consisting of a single tissue type. In order to achieve meaningful segmentation the size of the neighborhood must be large enough to contain more than one tissue class at any point tin the image. This should be considered when selecting the neighborhood size.
5.4.6 Segmentation Using Overlapping Mosaics
Yang et al. [35] developed a method to segment fat in MR images using overlapping mosaics. The segmentation technique consists of 3 steps:
(1)Mosaic bias field estimation
(2)Adipose tissue segmentation
(3)Consistency propagation
Following smoothening (low pass filtering) to remove noise the expression for the biased image in (5.2) becomes:
fbiased |
(x, y) = foriginal(x, y)β (x, y), |
(5.16) |
where fbiased(x, y) is the image after filtering. Assuming the bias field varies gradually across the entire image, log (β (x, y)) can be approximated by a piecewise
linear function. Yang divides fbiased(x, y) into a array of overlapping mosaics or subimages (Tij). Within each of the sub-images Log (β (x, y)) is assumed to be first
order linear, therefore, (x, y) Tij:
log f |
(x, y) = log( f |
original |
(x, y)) + aij |
x |
− |
x(0) |
+ bij y |
− |
y(0) |
+ cij, (5.17) |
biased |
|
|
|
ij |
|
ij |
|
where (x(ij0), y(ij0)) is the upper left voxel in each sub image. Optimal values for a and b are estimated by maximizing the function:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2 |
P = |
∑ |
∑ δ |
(log( f |
(x, y))) |
− |
aij |
x |
− |
x(0) |
− |
bij y |
− |
y(0) |
− |
ξ , |
|
biased |
|
|
|
ij |
|
ij |
|
ξ(x,y) Tij
(5.18) where δ (x) = 1 when x = 0, and ξ is the gray-scale intensity index of the image histogram. Cij is not calculated because it affects voxels in the sub–image uniformly, causing a change in position of gray–levels in the image histogram but not the shape. Once the image is corrected, the skewed and bimodal peaks discussed in Sect. 5.3.2 appear more distinctive.
When intensity inhomogeneities are corrected, image segmentation is carried out using a multi-level thresholding technique on each sub-image. Segmentation
108 |
D.P. Costello and P.A. Kenny |
Fig. 5.12 The intermediate processing results of the proposed algorithm applied to a synthetic phantom. (a) The original image with an SNR of 20 dB; (b) the optimum manually segmented result by using intensity thresholding; (c) the bias corrected image by using overlapping mosaics; (d) the result of initial segmentation; (e) the final fat distribution after applying intermosaic consistency propagation. This material is reproduced with kind permission from Springer Science+Business Media B.V [35]
is based on an automated thresholding technique that minimises the total variance within the three classes, fat soft tissue and background, giving 2 threshold values ξ1 and ξ2 [16].
A measure of confidence, λij, of the segmentation result is calculated, as not all sub images will contain all three tissue classes.
meanHij(ξ )
ξ ≥ξ2 |
|
|
λij = meanHij(ξ ) |
, |
(5.19) |
ξ <ξ1
Hij(ξ ) is the log transform of the image histogram. λij is likely to be large when all three tissue classes are present in the sub-image. However, when only one or two tissue classes are present λij will be much smaller indicating misclassification. The mosaic tile with the highest value of λij is used as a seed for consistency propagation. The regions of overlap between the seed tile and its nearest neighbors are compared. If any conflicting segmentation results are present, then the value for ξ2 in neighboring tile is changed to that of the seed. This process is propagated to all tiles within the image until segmentation result like those shown in Fig. 5.12 are achieved. Peng et al. [16], compared this technique to the gold standard, manual segmentation, and found that the mean percentage between the two was 1.5%.
5 Fat Segmentation in Magnetic Resonance Images |
109 |
5.5 Classification of Fat |
|
To appreciate the complexities associated with quantifying fat in medical images, it is important to know what exactly needs to be measured. The most common conflict in the literature is in the terminology used, (i.e. fat or adipose tissue) [52]. The difference between fat and adipose tissue is important when quantifying fat in MR images. Bone marrow in a typical T1w imaging sequence has the same graylevel as body fat. However, bone marrow adipose tissue is not classified as fat because it is connected to haematopoietic activity2 and not to obesity [53, 54]. Classification is further complicated by the subdivision of fat into three main categories: total body fat [15], visceral fat and subcutaneous fat [6]. Whole body fat includes the measurement of all adipose tissue except bone marrow and adipose tissue contained in the head, hands and feet [52]. A summary of the proposed classification of adipose tissue within the body is given by Shen et al. [52] and is presented in Table 5.1. Examination of body fat distribution involves the analysis of two or more of the fat categories outlined in Table 5.1.
Global segmentation algorithms such as thresholding require extra steps to classify fat. This can be achieved manually by drawing a region of interest around areas such as the viscera. Figure 5.13 shows the result of manual classification
Table 5.1 Proposed classification of total body adipose tissue as given by Shen et al. [52]
Adipose tissue compartment |
Definition |
Total adipose tissue |
Sum of adipose tissue, usually excluding bone marrow and |
|
adipose tissue in the head, hands, and feet |
Subcutaneous adipose tissue |
The layer found between the dermis and the aponeuroses |
|
and fasciae of the muscles. Includes mammary adipose |
|
tissue |
Superficial subcutaneous |
The layer found between the skin and a fascial plane in the |
adipose tissue |
lower trunk and gluteal-thigh area |
Deep subcutaneous adipose |
The layer found between the muscle fascia and a fascial |
tissue |
plane in the lower trunk and gluteal-thigh areas |
Internal adipose tissue |
Total adipose tissue minus subcutaneous adipose tissue |
Visceral adipose tissue |
Adipose tissue within the chest, abdomen, and pelvis |
Non-visceral internal adipose |
Internal adipose tissue minus visceral adipose tissue |
tissue |
|
Intramuscular adipose tissue |
Adipose tissue within a muscle (between fascicles) |
Perimuscular adipose tissue |
Adipose tissue inside the muscle fascia (deep fascia), |
|
excluding intramuscular adipose tissue |
Intermuscular adipose tissue |
Adipose tissue between muscles |
Paraosseal adipose tissue |
Adipose tissue in the interface between muscle and bone |
|
(e.g., paravertebral) |
Other non-visceral adipose |
Orbital adipose tissue; aberrant adipose tissue associated |
tissue |
with pathological conditions (e.g., lipoma) |
2Hematopoietic activity: pertaining to the formation of blood or blood cells.