- •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
168 |
L. Domanski et al. |
branches, their length, or the number of branching points. Branching structures typically possess a hierarchical organization, and it is important to capture and deliver information at each level. This higher order processing is described in Sect. 8.2.5.
There is renewed excitement nowadays around computational platforms such as Graphics Programming Units (GPUs) to speed up algorithms. This is particularly the case for high-throughput applications and for 3D and 4D datasets, where execution time still represents a major bottleneck. Section 8.3 describes the implementation of our linear feature detection algorithm on the GPU.
The technology presented in this contribution has been implemented in a WindowsTM package called HCA-Vision. For the user’s benefit, Sect. 8.4 outlines the software architecture and the main capabilities of this user-friendly tool (www.hca-vision.com).
Section 8.5 describes several representative applications in biological imaging. The first example demonstrates that even subtle phenotypic changes in the arbors of neurons in culture can be characterized if a sufficient number of neurons are analyzed. In this type of application, automation is central to reaching statistical significance. The importance of maintaining linear feature continuity is also highlighted by this example.
We have also used HCA-Vision with success to characterize the phenotype of astrocytes in response to drug-induced stress. In this application, fluorescently stained bundles of intermediate filaments were traced rather than neurites. Intermediate filaments are typical of the many polymers present within the cells, and we expect this application to grow in importance in the future.
In our final example, we describe how we used our sensitive linear feature detector to separate closely adjoining bacteria under minimal contrast, thus complementing the capability of more mainstream edge detectors.
Section 8.6 concludes this chapter by giving insight into recent developments, as well as by speculating about future directions.
8.2 Methods
There are many different algorithms suitable for linear feature detection. Good reviews are provided in [3,4]. Sun and Vallotton {Sun, 2009 #2}developed a fast linear feature detection method using multiple directional nonmaximum suppression (MDNMS). This approach is particularly intuitive and suitable for a nonspecialist audience. We will summarize the main steps of this algorithm in this section and describe its implementations on the GPU in Sect. 8.3.
8.2.1 Linear Feature Detection by MDNMS
Linear features are thin objects across which the image presents an intensity maximum in the direction of the largest variance, gradient, or surface curvature
8 High-Throughput Detection of Linear Features: Selected Applications... |
169 |
Fig. 8.1 Illustration of linear windows at four different directions. “x” indicates the centre of the linear windows. The other pixels in the window are shown as small circles. The length of the linear window is 7 pixels in this illustration
(i.e., perpendicular to the linear feature). This direction may be obtained by the use of computationally expensive Hessian-based detectors or by using matched or steerable filters. Rather than searching for the local direction of linear features, we use 1D nonmaximum suppression (NMS) [5] in multiple directions to identify candidate pixels on a linear feature.
NMS is the process of removing all pixels whose intensity is not the largest within a certain local neighborhood, that is searching for local maxima. The shape of this local neighborhood is usually a square or rectangular window. For our purpose, we choose this local neighborhood as an orientated linear window. A number of approaches are available to obtain the directional local maximum within a onedimensional window. For example, a local maximum can be obtained by using basic morphological operators to check whether a pixel value in the input image is equal to that in the dilated image [6].
Figure 8.1 shows four examples of linear windows at angles equal to 0, 45, 90, and 135◦. Additional directions, such as 22.5, 67.5, 112.5, and 157.5◦ can also be used. NMS is performed successively for each direction at every pixel. The result is independent of the order of the scanning process. The longer the linear window, the lower the number of candidate pixels detected in an image. The ideal scenario occurs when the linear window crosses the linear feature at right angles as this will produce maximum contrast.
The outputs of the directional NMS are binary images, as opposed to grayscale images produced by most directional filter methods [3]. This eliminates the need for an arbitrary thresholding step. Linear features are detected by the union of MDNMS responses, that is we combine the NMS responses at each orientation.
The algorithms for linear feature detection using NMS can be easily extended for 3D images by using 3D linear windows. We used either 3 or 9 directions in 3D, although additional directions may be necessary, depending on the data.
170 |
L. Domanski et al. |
Fig. 8.2 Cross section through a linear feature
8.2.2 Check Intensities Within 1D Window
Bona fide linear features are characterized by an approximately symmetric intensity profile across the feature, as opposed to edges which show a step-like profile. This translates into approximately equal intensity values around the central pixel on both sides of the linear window. Figure 8.2 illustrates the parameters that characterise the shape of a profile. Imax is the value of a local directional maximum at the centre pixel of the linear window. Iaverage1 and Iaverage2 are the average intensity values over the two sides of the local maximum within the local window; and Idiff1 and Idiff2 are the differences between the maximum value and these two average values. For a genuine linear feature, both Idiff1 and Idiff2 should be large. The parameter Idiff1 and Idiff2 can be used to control the sensitivity of the algorithm. Lowering the values of Idiff1 and Idiff2 increases the number of linear features detected.
8.2.3 Finding Features Next to Each Other
Some images present linear features in close proximity to one another. The NMS process will detect only one feature in each linear window. Figure 8.3 shows an example where four local maxima lie within the linear window. We can extend the NMS process so that multiple local maxima are detected for any particular linear window. Once a local maximum is found at the center of a linear window, a second local maximum search is performed within the original window using a smaller window size. The use of a smaller window size allows additional local maxima to be found within the original window. The newly found local maxima are ordered based on their maximum intensity as shown in Fig. 8.3. To prevent detecting false peaks due to noise, we also check that the intensity does not deviate significantly from Imax and that it conforms to the symmetry condition defined in Sect. 8.2.2. For example, in Fig. 8.3, Imax 3 and Imax 4 are two local maxima, but they are not strong enough to be retained.
8 High-Throughput Detection of Linear Features: Selected Applications... |
171 |
Fig. 8.3 Multiple local maxima within a linear window
Fig. 8.4 Endpoint (red) and its neighborhood (green) for gap linking. The shortest path is shown in blue
8.2.4 Gap Linking for Linear Features
The union of the multiple responses of NMS at different directions generates many small objects, which may not belong to genuine linear features. We fit an ellipse to each object (set of connected pixels) and remove them if the major axis is too small. Alternatively, a simple pixel count can be used.
The continuity of the linear features may be broken due to noise or weakness of linear features. This results in small gaps in the combined responses of NMS. We restore continuity by joining linear feature endpoints to neighboring linear features through a shortest path. Here, a shortest path should be understood as a connection, which displays a high average intensity along its length. A standard technique called dynamic programming limits the combinatorial explosion that would otherwise be associated with the exploration of all candidate paths [7]. The computational overhead also remains small because the operation is only performed on small gaps – typically below 20 pixels in length. Figure 8.4 illustrates the process of gap linking from an endpoint to the skeleton in its neighborhood. Note that the domain shown in green in Fig. 8.4 is first transformed into polar coordinates prior to calculating the shortest path.