KTH, Division of Neuronic Engineering, have a finite element model of the head. However, this model does not contain detailed modeling of the brain. This thesis project consists of finding a method to extract brain tissues from T1-weighted MR images of the head. The method should be automatic to be suitable for patient individual modeling. A summary of the most common segmentation methods is presented and one of the methods is implemented. The implemented method is based on the assumption that the probability density function (pdf) of an MR image can be described by parametric models…
Contents
1. Introduction
1.1. Background
1.2. Assignment
1.3. Outline
2. Segmentation, Classification, and Labeling
3. MR Images
3.1. Weighting
3.1.1. T1-weighted Images
3.1.2. T2-weighted Images
3.2. Patient Dependency
3.3. Artifacts
3.3.1. Intensity Inhomogeneities
3.3.2. The Partial Volume Effect
4. Segmentation Methods
4.1. Features
4.2. Manual Segmentation
4.3. Thresholding
4.4. Atlas-Based Methods
4.5. Watershed
4.6. Region Growing
4.7. Active Contours
4.8. Classifiers
4.9. Clustering
4.10. Markov Random Fields
4.11. Fuzzy Connectedness
5. Why an Automatic Method?
6. The Chosen Method
7. Theory for the Implemented Method
7.1. Preprocessing
7.1.1. Removal of Background Voxels
7.2. Modeling the Intensities Emitted by One Tissue
7.3. Modeling the Intensity Distribution
7.4. Voxel Classification
7.4.1. Soft Classification
7.4.2. Estimation of the Gaussian Parameters
7.4.3. Estimation of the Bias Field
7.4.4. The Expectation-Maximization Algorithm
7.4.5. Hard Classification
7.5. Segmentation of the Brain Using Morphological Operations
7.5.1. Dilation
7.5.2. Erosion
7.5.3. Opening
7.5.4. Closing
8. Implementation
8.1. Preprocessing
8.1.1. Removal of Background Voxels
8.1.3. Initialization of the EM Algorithm
8.2. Voxel Classification
8.2.1. Modeling the Bias Field
8.2.2. Spatial Filtering of the Classifications
8.3. Segmentation of the Brain Using Morphological Operations
9. Results
9.1. Synthetic Images
9.1.1. Bias Correction
9.2. MR Images
9.2.1. Removal of Background Voxels
9.2.2. Voxel Classification
9.2.4. The Number of Classes
9.2.5. Initialization of the EM Algorithm
9.2.6. Bias Correction
9.2.7. Spatial Filtering of the Soft Classifications
9.3. Segmentation of the Brain
10. Discussion
10.1. Preprocessing
10.1.1. Logarithmic Transformation of the Intensities
10.1.2. Initialization of the EM algorithm
10.1.3. The Number of Classes
10.2. Voxel Classification
10.2.1. Spatial Considerations
10.2.2. Bias Correction
10.3. Segmentation of the Brain
10.4. Extensions of the Method
10.5. Validation
11. Conclusions
12. References
Author: Caesar, Jenny
Source: Linköping University
Download URL 2: Visit Now