UNIVERSITITEKNOLOGI MARA
METACARPAL PHANTOM RADIOGRAPH EDGE DETECTION USING GENETIC ALGORITHM
GRADIENT BASED GENOTYPE
NORHARYATIBINTIMD ARIFF
Thesis submitted in fulfillment of the requirements for Bachelor of Science (Hons) Intelligent System
Faculty of Information Technology And Quantitative Sciences
MAY 2007
APPROVAL
Metacarpal Phantom Radiograph Edge Detection Using Genetic Algorithm Gradient Based Genotype
BY
NORHARYATIBWU MD ARIFF
This thesis was prepared under the direction of thesis supervisor. Madam Noor Elaiza Binti Abdul Khalid. It was submitted to the Faculty of Information Technology and Quantitative Sciences and was accepted in partial fiilfllhnent of the requirements for the degree of Bachelor of Science Honors Intelligent System.
Approved by:
Date: 31 May 2007 Madam Noor Elaiza Binti Abdul Khalid Thesis Supervisor
n
DECLARATION
I certify that this thesis and the research to which it refers are the product of my own work and that any ideas or quotation from the work of other people, published or otherwise are fully acknowledged in accordance with the standard referring practices of the discipline.
MAY 4,2007 NORHARYATIBINTIMD ARIFF 2004107457
111
ABSTRACT
The conventional criterion for fi"acture risk assessment is measured based on bone mineral density (BMD) value that is produce by X-ray. Even if there is a strong association between bone strength and BMD, nowadays it is well accepted that this is not sufficiently reliable predictor of fi^cture risk in osteoporotic patients. Therefore, there is a growing need for better predictor of bone strength. The image processing is one of the methods to measure the bone strength. It is need to get the optimal result of osteoperosis or osteopenia detection. Using image segmentation, the bone strength can be measure by looking for the great image on the length of the outline cortical.
For the image segmentation, genetic algorithm are use to segment the bone image and it is a new method that applies in the image processing field. Genetic algorithms have several steps. For the initial population, 200 pixels will be taken to be the initial population in randomly. From the initial population, the fitaess fimction has to calculate to get the fittest pixels. Fitaess fimction is calculated based on the gradient which is the length fi*om each pixel. The range of pixel value and the position between the characteristic of the bone are define. The fittest pixel values that are fit with the characteristic are selected to make the crossover and mutation. After that, the next generation will be process until it get the satisfy value to get the optimum line between the cortical bone and trabecular bone. The objectives has been achieved and found the outline of the cortical bone. The results come out with the mean and standard deviation of the cortical bone length. The prototype is capable to shows the outline of the cortical bone and the accuracy of the outline is 30% to 50% after make the comparison fi-om other researcher.
TABLE CONTENTS
CONTENT PAGE APPROVAL ii DECLARATION iii ACKNOWLEDGEMENT iv
ABSTRACT v TABLE OF CONTENTS vi
LIST OF FIGURES xi CHAPTER ONE: INTRODUCTION
1.0 Background 1 1.1 Problem Statement 1 1.2 Objective of the project 2 1.3 Scope of the project 2 1.4 Significance of the project 2 1.5 Expected Outcomes/Deliverables 2
1.6 Summary 3
CHAPTER TWO: OSTEOPOROSIS AND BONE ANATOMY
2.0 Introduction 4 2.1 Osteoporosis 4 2.2 Bone Anatomy 6 2.3 Normal Bone Remodelling Process 8
2.3.1 Bone Remodelling 8 2.3.2 Bone Resorption 9 2.3.3 Bone Formation 9 2.4 Bone Strength and Bone Mineral Density 10
2.5 The Symptoms of osteoporosis 10 2.6 The consequences of osteoporosis 11 2.7 The risk factors of osteoperosis 12
2.8 Summary 13
VI