Segmentation of Hand Bone for Bone Age Assessment (eBook)

Dr. Hum Yan Chai received his PhD in the field of Biomedical Imaging from the Universiti Teknologi Malaysia (UTM). He is currently working as a researcher in the Medical Implant Technology Group (MEDITEG), Materials and Manufacturing Research Alliance. His research interests are: computerized bone age assessment, Digital X-Ray imaging, medical image processing, filter design, fuzzy logic, medical computing and performance optimization. He serves as the member of editorial board for a few international journals in relevant fields. He is also a program committee member and peer reviewer for over 15 international conferences.

1     Introduction1.1  Introduction1.2  Background of the problem1.3  Problem statements1.4  Objectives of the book/brief1.5   Scopes1.6   Provided information and insights1.7   Book Organization2     The conventional segmentation methods2.1   Introduction2.2   Thresholding     2.2.1  Global Thresholding     2.2.2  Adaptive Thresholding     2.2.3  Dynamic Thresholding     2.2.4  Automated Thresholding     2.2.5  Summary2.3   Edge-based      2.3.1  Edge Detectors      2.3.2  Edge linking      2.3.3  Summary2.4   Region-based       2.4.1  Seeded Region Growing       2.4.2  Region Splitting and Merging       2.4.3  Summary3    The advanced segmentation method3.1  Hybrid-based     3.1.1  Watershed Segmentation     3.1.2  Summary 3.2   Deformable Model      3.2.1 Active Contour Model      3.2.2 Active Shape Model      3.2.3 Active Appearance Model      3.2.4 Summary4    The possible solution 4.1  Introduction4.2  The Proposed Segmentation Framework4.3   Pre-processing      4.3.1  The  Proposed MBOBHE             4.3.1.1 Modeling of  Criteria as Single Modal Objective Beta Function              4.3.1.2 Optimal Solution of the Aggregated Multiple Objectives Function               4.3.1.3 Histogram Decomposition               4.3.1.4 Execution of GHE on Each Sub-Histogram         4.3.2  The Application of Anisotropic Diffusion                4.3.2.1 Parameter-free Diffusion Strength Function                4.3.2.2 Automated Scale Selection4.4   The Proposed Adaptive Crossed Reconstruction (ACR) Algorithm Design       4.4.1  Clustering Algorithm Applied in the Proposed Segmentation Framework       4.4.2  Automated Block Division Scheme in Adaptive Segmentation                4.4.2.1 The Framework of the Proposed Scheme                4.4.2.2 The Mechanism of the Automated Fuzzy Quadruple Division Scheme4.5  Quality Assurance Process       4.5.1  Gray Level Intensity of Interest Identification for Elimination         4.5.2  Hand Bone Edge Detection Technique Using Entropy         4.5.3  The Area Restoration and Elimination Analysis4.6   Summary5    Result analysis and discussion5.1  Introduction5.2  Performance Evaluation of the Proposed MBOBHE5.3  Anisotropic Diffusion in the Proposed Segmentation Framework5.4  Segmentation Evaluation      5.4.1  User-specified parameters               5.4.1.1  Active Appearance Model               5.4.1.2  The Proposed Framework                5.4.1.3  Interpretation       5.4.2  Segmentation Accuracy               5.4.2.1 Evaluation on Automated Fuzzy Quadruple Division Scheme                5.4.2.2 Evaluation on Quality Assurance Process               5.4.2.3 Accuracy Evaluation of the Proposed Segmentation Framework5.5  Summary6 Conclusion and Recommendation6.1 Conclusion6.2  Future works