Next:
List of Figures
Up:
UNIVERSITAT DE GIRONA Department
Previous:
Abstract
Contents
List of Figures
List of Tables
Introduction
Breast Cancer
Mammography
Mammographic Abnormalities
Digital Mammography
Image Acquisition
Image Storage
Image Display
Computer-Aided Systems
The Benefits of CAD
Commercial CADs
Scope of the Research
HRIMAC Project
Objectives of the Thesis
Thesis Outline
A New Framework for Mass Detection
Document Overview
A Review of Automatic Mass Segmentation Techniques
Introduction
Breast Profile Segmentation
Mass Segmentation Using One Single View
Region-Based Methods
Contour-Based Methods
Clustering and Thresholding Methods
Model-Based Methods
Mass Segmentation Using Two or More Images
Comparison of Left and Right Mammograms
Comparison of Two Mammographic Views
Temporal Comparison of Mammograms
Evaluation of Mass Segmentation Methods
Evaluated Mass Segmentation Methods
Evaluation Methodology: ROC and FROC Curves
Mass Segmentation Results
Discussion
Breast Density Classification
Introduction
A Survey on Automatic Breast Density Classification
A New Proposal for Automatic Breast Density Classification
Finding Regions with Similar Tissue
Extracted Features
Classification
Results
MIAS Database
DDSM Database
The Importance of the Segmentation Step
Discussion
Comparison with the Works which Classifies into BIRADS Categories
Conclusions
Mass Segmentation Using Shape and Size Lesion Information
Introduction
A Brief Review on Deformable Template Models
From Eigenfaces to Eigenmasses
Eigenfaces
Eigenmasses and Eigenrois
Probabilistic Mass Contour Template
Template Based Detection
Prior Distribution
Likelihood
Posterior Probability Density
Final Considerations
Results
MIAS Database
Málaga Database
Discussion
False Positive Reduction
Introduction
PCA-Based False Positive Reduction
2DPCA-Based False Positive Reduction
Evaluation of the False Positive Approaches
MIAS Database
DDSM Database
Combining the Bayesian Pattern Matching and the False Positive Reduction Step
MIAS Database
Training and Testing using Different Databases
Computational Cost
Discussion
Automatic Mass Segmentation using Breast Density Information
Introduction
Including Breast Density Information in our Mass Detection Approach
Results Obtained Including Breast Tissue Information
False Positive Reduction Step with Breast Density Information
Comparison of the Method with Existing Approaches
Testing the Approach Using a Full-Field Digital Mammographic Database
Discussion
Conclusions
Summary of the Thesis
Contributions
Further Work
Increasing the Reliability of the Proposal
Future Research Lines Departing from this Thesis
Technological Further Work
Related Publications
Breast Profile Segmentation
Introduction
A Fast Breast Segmentation Algorithm with Pectoral Muscle Suppression
A Contour-Based Approach to Breast Skin-Line Segementation
Skin-Line Detection in Scale Space
Seed Point
Contour Growing
A Brief Description of the Used Mammographic Databases
Introduction
MIAS
Database Characteristics
Used Mammograms Containing Masses
DDSM
Database Characteristics
Used Mammograms Containing Masses
Málaga Database
Database Characteristics
Used Mammograms Containing Masses
Trueta Database
Image Characteristics
Used Mammograms Containing Masses
Evaluation Methodologies
Introduction
Evaluation of Classifiers
Confusion Matrices
ROC Analysis
Detection Evaluation
ROC Analysis
FROC Analysis
Bibliography
Arnau Oliver 2008-06-17