i ACKNOWLEGDEMENT In the name of Allah SWT ...

ACKNOWLEGDEMENT

In the name of Allah SWT the most Gracious and the most Merciful. Alhamdulillah for giving me the strength and the ability to complete the final year project for the Bachelor of Engineering (Biomedical Electronic Engineering) at Universiti Malaysia Perlis (UniMAP).

First of all, I would like to take this opportunity to express my most profound gratitude and appreciation to my supervisor Dr. KHAIRUL SALEH BIN BASARUDDIN for his constant advice, kind guidance, valuable ideas, suggestions and generous support throughout the entire project. You been very helpful and understandable, without your assistance and guidance, this project will not be successful. Your supervision enlightened my way to do this project. Your guidance is fully appreciated with warm heart. Thank you very much. I also wish to thank a donor’s family for donating by a male cadaver aged 69 years old which our sample of study the fourth lumbar (L4) vertebral trabecular bone was extracted from it.

I would like to thank School of Mechatronics Engineering for providing me a good research environment to complete this project. I would also like to thank my lecturers for their support, advice, and their readiness to give their helping hands for me to finish my final year project successfully.

Finally, I would like to thank my family who has always stood by my side and gives me countless support to complete this project. They were always the back bone for me and be a pillar for me in completing my final year project.

i

APPROVAL AND DECLARATION

This project is titled Determination Of Mechanical Properties For Cancellous Bone Based On Image Greyscale Value was fully prepared and submitted by YOUSIF ALI YOUSIF ALGABRI (Matriculation Number: 111150001-5) and has been found satisfactory in terms of scope, quality and presentation as a partial fulfillment of the requirement for the Bachelor of Engineering (Biomedical Electronic Engineering) in Universiti Malaysia Perlis (UniMAP).

Checked and Approved by

(Dr. KHAIRUL SALEH BIN BASARUDDIN) Project Supervisor

School of Mechatronic Engineering Universiti Malaysia Perlis

MAY 2015

ii

PENENTUAN SIFAT MEKANIK UNTUK CANCELLOUS BONE BERDASARKAN IMEJ SKALA KELABU NILAI

ABSTRAK

Projek ini membentangkan penentuan sifat mekanik untuk tulang cancellous berdasarkan nilai skala kelabu. Pada tahun-tahun kebelakangan ini, peningkatan pesakit osteoporosis menjadi beban kepada hospital, oleh itu ia adalah satu isu yang mendesak untuk mencari pencegahan dan rawatan kaedah-kaedahnya. Sejak tulang cancellous adalah metabolisme lebih aktif daripada tulang kortikal, tulang cancellous sering digunakan untuk diagnosis osteoporosis dan telah mendapat perhatian dalam kajian tulang. Projek ini juga dijalankan untuk mencadangkan satu skim pengiraan yang akan dapat meramalkan berkesan modulus elastik jelas tulang trabekular memandangkan ketidaktentuan yang terutamanya disebabkan oleh orientasi pemodelan dan trabekular kekakuan berasaskan imej. Kesan daripada pemodelan berasaskan imej yang memberi tumpuan kepada hubungan itu juga disiasat. Selain itu, kaedah berangka untuk menjelaskan morfologi seni bina rangkaian trabekular kompleks di lumbar tulang belakang manusia melalui teknik pasca pemprosesan baru untuk tekanan dikira mikroskopik dengan kaedah penyeragaman itu. Micro-CT berdasarkan imej-teknik pemodelan digunakan dan kaedah-hati tetapi intuitif dan mudah untuk digunakan untuk model mikrostruktur. Sifat homogenized makroskopik yang termasuk bukan sahaja modulus Young tetapi juga modulus ricih boleh menjelaskan kekukuhan tulang yang sihat. Nilai utama jelas modulus Young berkesan yang diramalkan dalam paksi utama ditemui pada kirakira 159,45 MPa dan nisbah 0,1488 Poisson.

iii

DETERMINATION OF MECHANICAL PROPERTIES FOR CANCELLOUS BONE BASED ON IMAGE GREYSCALE VALUE

ABSTRACT

This project presents determination of mechanical properties for cancellous bone based on greyscale value. In the recent years, the increase of patients with osteoporosis is becoming burden to hospitals, thus it is an urgent issue to find its prevention and treatment methods. Since cancellous bone is metabolically more active than cortical bone, cancellous bone is often used for diagnosis of osteoporosis and has received much attention within the study of bone. This project was also undertaken in order to propose a computational scheme that will be able to predict the effective apparent elastic moduli of trabecular bone considering the uncertainties that are primarily caused by image-based modelling and trabecular stiffness orientation. The effect of image-based modelling which focused on the connectivity was also investigated. Moreover, a numerical methodology to clarify the morphology of complex trabecular network architecture in human lumbar vertebra by means of the new post-processing technique for calculated microscopic stress by the homogenization method. Micro-CT image-based modeling technique is used and careful but intuitive and easy-to-use method for microstructure model. The macroscopic homogenized properties that include not only the Young’s moduli but also shear moduli could explain the stiffness of healthy bone. The main value of the predicted effective apparent Young’s moduli in principal axis was found at approximately 159.45 MPa and 0.1488 Poisson’s ratio.

iv

TABLE OF CONTENTS

ACKNOWLEGDEMENT

i

APPROVAL AND DECLARATION

ii

ABSTRAK

iii

ABSTRACT

iv

TABLE OF CONTENTS

v

LIST OF FIGURES

viii

LIST OF TABLES

x

CHAPTER 1 INTRODUCTION

1

1.1 Background

1

1.2 Problem statements

2

1.3 Objectives

3

1.4 Scopes

3

1.5 Thesis outlines

4

CHAPTER 2 LITERATURE REVIEW

5 5

2.1

Introduction

5

2.2 Bone structure

5

2.3 Mechanical properties of bone

6

2.3.1

Shearing Stresses and Strains

7

2.3.2

Poisson ratio

8

2.3.3

Modulus of elasticity

9

2.4

Image processing

2.4.1

12

Image segmentation

12

2.4.2 Image Thresholding

12

2.4.3

14

Single Thresholding

v

2.5

Finite element analysis of bone model

15

2.5.1 Background

15

2.5.2 Development of FE bone model

15

2.6

Homogenization theory

16

2.7

Morphology of cancellous bone

17

2.8

Literature reviews summary

19

CHAPTER 3 METHODOLOGY

21

3.1

Introduction

21

3.2

Image processing

21

3.2.1

CT-image acquisition

21

3.2.2

Image segmentation procedure

23

(a)

Averaging filter

23

(b)

Superimpose theory

23

(c)

Image contrast technique

24

3.3 Threshold the Image

24

3.4

25

Development of 3D finite element model of cancellous bone.

3.4.1

Bone tissue properties

26

3.4.2

Measurement of trabecular morphology for mechanical properties

26

3.4.3

Computational analysis

27

3.5

Homogenization method

28

3.6

Normal distribution graph

28

CHAPTER 4 RESULTS AND DISCUSSION

30

4.1 Introduction

30

4.2 Image segmentation procedure

30

4.3 Geometrical models

34

4.4 Effect of image greyscale value on mechanical properties of cancellous bone

36

4.4.1 Young’s moduli

36

4.4.2 Poisson’s ratio

38

4.4.3 Shear moduli

40 vi

4.5 Effect homogenized trabecular morphology on mechanical properties of cancellous bone

43

4.6 Prediction of effective homogenized mechanical properties.

44

4.7 Discussion

46

CHAPTER 5 CONCLUSION

47

5.1 Summary

47

5.2 Recommendation

48

REFFERENCES

49

APPENDIX A: SOURCE CODE FOR IMAGE CROPPING

53

APPENDIX B: SOURCE CODE FOR IMAGE THRESHOLD

55

APPENDIX C: BONE VOLUME CALCULTION

56

vii

LIST OF FIGURES

Figure 1.1 Longitudinal section of the humerus (upper arm bone), showing outer compact bone and inner cancellous (spongy) bone

2

Figure 2.1 a) Photo of a cross section of elk antler, (b) Longitudinal section of a femur

6

Figure 2.2 The engineering stress-strain diagram.

8

Figure 2.3 Original image (a) and result of segmentation (b).

13

Figure 2.4 Measurement of Tb morphology for 3D FE model

18

Figure 3.1 flowchart for the overall project.

22

Figure 3.2 Measurement of Tb morphology for 3D FE model

26

Figure 3.3 Normal distribution graph

29

Figure 4.1 Original CT-image for cancellous bone

31

Figure 4.2 The original image after applying the crop process

31

Figure 4.3 The original cropped after applying filtering (fspecial) command by MATLAB software

32

Figure 4.4 After applying superimpose technique and image contrast

32

Figure 4.5 And then after thresholding method procedure

33

Figure 4.6 Model 1 for threshold Tlower value

34

Figure 4.7 Model 2 for threshold Tmean value

35

Figure 4.8 Model 3 for threshold Tupper value

35

viii

Figure 4.9 The effect of threshold on Young's moduli in axis E11

36

Figure 4.10 The effect of threshold on Young's moduli in axis-1, E22

37

Figure 4.11 The effect of threshold on Young's moduli in axis E33

37

Figure 4.12 The effect of threshold on Poisson's ratio in axis 12

38


39


39

Figure 4.15 The effect of threshold on shear moduli in axis G12

40


41


41

Figure 4.18 The effect of threshold on morphology of cancellous bone

44

ix

LIST OF TABLES

Table 2.1 A comparison of elastic moduli of trabecular tissue, Etissue (mean±SD, in GPa), Bayraktar et al

11

Table 2.2 Measured values for the elastic modulus of single unmachined trabeculae and of trabecular material summarized by Lucchinetti et al

11

Table 2.3 Summary of literature review

19

Table 4.1 The result for the morphology of cancellous bone

43

Table 4.2 The predicted results for homogenization property

45

x

CHAPTER 1

INTRODUCTION

1.1

Background The human bones mechanical properties (elastic moduli and strength) and their own

loading situations determine the fractures bone risks. In case for the elderly, as fractures of bone are catastrophic events, lead to increasing the rate of mortality and decreasing mobility. The determination of bones mechanical properties can be obtained at micro level (consist of cancellous and cortical bone) and micro level (cancellous bone region). The mechanical properties cancellous bone regions are of certain importance since the most of the bone osteoporotic fractures are initiated in those regions. Image processing techniques will be integrated with multiscale finite element method in this project to determine the reliable mechanical properties of cancelous bone. In addition, cancellous bone makes up about 20 percent of the human skeleton, providing structural support and flexibility without the weight of compact bone.[1] It is found in most areas of bone that are not subject to great mechanical stress. It makes up much of the enlarged ends (epiphyses) of the long bones and is the major component of the ribs, the shoulder blades, the flat bones of the skull, and a variety of short, flat bones elsewhere in the skeleton. Cancellous bone is usually surrounded by a shell of compact bone, which provides greater strength and rigidity. Cancellous bone can develop into compact bone through the action of bone-forming cells called osteoblasts. It is in this manner that all long bones develop in the embryo. The osteoblasts deposit new bone matrix in layers around the trabeculae, which thus enlarge at the expense of the spaces between them. Eventually the spaces are eliminated, and immature compact bone is produced. See also bone formation as shown in Figure 1.1. [1] 1

Figure 1.1: Longitudinal section of the humerus (upper arm bone), showing outer compact bone and inner cancellous (spongy) bone [1]

1.2

Problem statements The reconstruction of image-based finite element model using a single threshold value

could jeopardize the numerical accuracy because some of the elements might be eliminated due to inaccuracies capturing the medical image. In addition, by considering multiple threshold values in determining the segmentation image will be able to improve the reliability of the reconstructed image-based finite element model. Hence, this project will be undertaken to determine the mechanical properties for cancellous bone using a new technique that utilizing the image grey scale value and multiscale finite element method. This project will also investigate the effect of threshold value on determination of material properties for cancellous bone. Therefore, the aim of 2

this project is to determine and investigate the mechanical properties for cancellous bone based on greyscale value which will be captured using micro-CT scanner.

1.3

Objectives

This study embarks on the following objectives: i.

To suggest an appropriate procedure for segmentation of cancellous bone image.

ii.

To investigate the effect of threshold value on the homogenized mechanical properties and trabecular morphology of cancellous bone.

iii.

To predict the effective homogenized mechanical properties of cancellous bone based on probability function.

1.4

Scopes This project will focus on the determination of mechanical properties for cancellous bone

by simulation utilizing the image processing technique and finite element method. The mechanical properties to be investigated in this study are the common elastic constants, which are Young's moduli, Poisson’s ratio, and shear moduli. Moreover, the image will be segmented by using image threshold technique. To add, the obtained threshold values have a major effect on the mechanical properties in term of determining the stiffness of the healthy bone and on the trabecular morphology parameters measurements. Certainty, the effective homogenized mechanical properties of cancellous bone are predicted that conducted as linear static analysis with periodic boundary conditions for probability function. The model of cancellous bone is based on image greyscale value of the threshold. Therefore, the analysis will be conducted as linear static analysis with periodic boundary conditions.

3

1.5

Thesis outlines This thesis includes five chapters, the first chapter discuses about the background of

mechanical properties of bone, cancellous bone, and image segmentation. In addition, problem statement, objectives and project scope are also presented in this chapter. The second chapter is the studies from research papers and journals of other researchers that relates to the project are presented. The third chapter is the Methodology. This chapter is divided into four sub-chapters; data acquisition stage, data pre-processing, feature extraction and classification stag. The fourth chapter is result and discussion. In this chapter, outcome of the project is displayed and discussed. It includes brief limitation throughout the progress of the project. The fifth chapter is a conclusion. Here, the overall project is bring about to a conclusion. The chapter includes future development and recommendations proposal for the project.

4

CHAPTER 2

LITERATURE REVIEW

2.1

Introduction

This chapter includes the studies, research and experiments that have done before and have strong relation to the study of this project. In the beginning, bone structure has been explained clearly and the significant structure of human bone from a trusted sources and researches. In addition, a brief on the mechanical properties of bone and stating the most common properties Young’s modulus , Poisson’s ratio and stress-strain property. Moreover, the general principles of image processing such image segmentation and image thresholding. The last part of chapter two is about the finite element model and homogenization theory with implementing the table for literature review summary.

2.2

Bone structure The bone is a polymer and ceramic composite, polymer as collagen and hydroxyapatite or

calcium phosphate for ceramic. Bone is the main component for our body structure. Of course, the bone has other several functions as well. However, we will focus here on the mechanical properties and its performances. Cancellous bone exists in bones flat and core bone while the cortical is can be found in the longer bones such as fibula, femur, tibia and etc. [1] 5

Figure 2.1(a) shows antler of deer cross-sectional picture. Porosity decreases bone strength and antler, it also decreases the weight. That can be formed in manners that are provided strength just when it is needed. The cancellous bone porosity is providing appropriate mechanical properties. The porosity and its way of structuring can determine the mechanical strength. Also the pores are performing different physiological functions, and containing the bone marrow [1]. Nevertheless Figure 2.1(b) shows the cross sectional view of long bone structure. The outer part is cortical bone that is compact and the inner part is called cancellous bone that is more spongelike and porous.

(a)

(b)

Figure 2.1: a) Photo of a cross section of elk antler, (b) Longitudinal section of a femur [1]

2.3

Mechanical properties of bone Initially, every normal bone has mechanical properties such as elastic moduli and strength

with own loading states determine the bone fracture risks. In certain for elderly, such these fractures of bone are catastrophic events, which leading to an increasing of the rate of mortality and a decreasing for mobility. A non-invasive and accurate assessment of mechanical properties of the bones could be significant to the risk of bone fractures, such as prevention of measures could 6

be observe in time. The determination of bones mechanical properties can be declared by either the internal or external architecture of the bones and the properties of tissue. Thus, there is a possibility that the mechanical properties of the bone can have a good prediction to be achieved by the consideration of the aspects of architectural. At the recent time, to account for every aspect of the internal architectural of the bone, the new techniques of microfinite element (micro-FE) are introduced. These methods are based on high resolution 3D bone images specimens, captured by using micro-CT or other imaging techniques with high resolution (10-50 µm for such methods is a typical resolution) and reconstructed in the computer. The cancellous bone internal architecture in details can be represented by the model of micro-FE after converting this reconstruction. The specimens elastic properties as a whole can be calculated when the FE problem solved for different loading terms [1, 2]. The mechanical properties of cancellous bone are depending on orientation or quantities of trabecular bone [3, 4]. The mechanical properties of cancellous bone are supposed to be constant with no relying on structural arrangement and parts. Moreover like this assumption not established in the studies of trabecular bone. In addition, the not enough shape assessment caused a big margin of errors in modulus of elasticity [5, 6].

2.3.1 Shearing Stresses and Strains The deformation is not all elongational or compressive, where the concept of strain needs to be extended to be included the effects of shearing or distortion the shearing distortions nature can be illustrated, the square inscribed grid on the specimen of tensile. The stress-strain curve is significant graphic measure the mechanical properties of materials. As known, the extremely important mechanical response material test is the tensile test, in which one end of a rod or wire specimen is clamped in a loading frame and the other subjected to a controlled displacement δ. 7

Therefore, the stress (σe) plots against the strain (e), the stress-strain engineering curve as shown in Figure 2.2.

Figure 2.2: The engineering stress-strain diagram. [5, 6]

2.3.2 Poisson ratio In general, if there are two different directions, the tensile stress with positivity is contributing a negative compressive strain from one direction to another. For instance, when stretching the rubber band for making it longer in one direction at the same time it makes it thinner in the other directions. The Poisson effect called a lateral contraction along the longitudinal effect, Poisson's ratio is a fundamental property known as: [6]

2.1

8

While the negative sign in the formula represent the sign of change between the longitudinal and lateral strain. Moreover, few studies describe experimental determination of Poisson's ratios of cortical bone. Reilly and Burstein assumed transverse isotropy of fibro lamellar bone, and used extensometers to measure strains in two orthogonal directions concurrently. They found Poisson's ratio values which ranged between 0.29 and 0.63. Ashman et al [7] reported on the use of an ultrasonic continuous wave technique, and found Poisson's ratio values which ranged between 0.27 and 0.45. Pithioux et al also used an ultrasonic method, and found Poisson's ratios between 0.12 and 0.29. Despite this wide range of reported values (0.12–0.63), many studies, especially finite element analyses, often use values in the much narrower range of 0.28–0.33.[7]

2.3.3

Modulus of elasticity Modulus of elasticity is known also as the Young’s modulus or tensile modulus; it is a

method to measure the elastic of material stiffness and also is a quantity can be used to identify the materials. The stiffness distinguish is very important, it is a measurement of the needed loading to bring about the deformation of the material contained in the force, this is usually indicates to the resistance of the materials to break failure or excessive deformation. Generally, the stiffness measured through the relatively application compact loads, much less than the fracture, and the result deformation measuring. This relationship, commonly known as Hooke's law, is written algebraically as: [2] 2.2 Where  is the force applied,  is a constant of proportionality called the stiffness and having units of lb/in or N/m and  is the extension of stiffness spring(m). Stiffness as specified by k is not only the function of the material, but it is also affected the shape of the sample [2].

9

2.3

Where  is young’s moduli,  is the change in stress and  is strain change. By comparing experimental data from uni-axial tension and compression of human femoral neck using FE calculation, Bayraktar and Keaveny concluded that the uniformity of apparent yield strains is primarily the result of the highly oriented architecture that minimizes bending as shown in Table 2.1. [8] Furthermore,

Lucchinetti et al summarized the potentials and limitations of micro-

mechanical testing of bone trabecular [8]. The range of the Young’s modulus given in this review, is 1 GPa up to 15 GPa see Table 2.2. This wide range was explained to be due to difficult sample preparation, handling and testing and furthermore due to the anisotropic and inhomogeneous bone material First micro-bending tests on single trabecular using a 3-point bending test, were performed by Lucchinetti [8].The measured deflection of a single trabecular under a load of 100 mN was about 1.1 µm. In a review, Lucchinetti et al (2000) stated that an error of 10 % in the surface geometry, mainly the variation of the thickness, is amplified to a 40 % error of the elastic modulus. [8]

10

Table 2.1: A comparison of elastic moduli of trabecular tissue, Etissue (mean±SD, in GPa), Bayraktar et al [8]

Table 2.2: Measured values for the elastic modulus of single unmachined trabeculae and of trabecular material summarized by Lucchinetti et al [8].

11

2.4

Image processing

2.4.1 Image segmentation

The segmentation is known as dividing the image into parts and also adding a structure to the raw image. In medicine cases, this is involving in determining which parts of the image is a tumor, or the separation of the white matter of the gray matter in the brain scan [9]. Image segmentation is a distinguishing process between background and objects of an image. It is the task of the basic processing for several applications that rely on computer vision, such as medical imaging, and locate objects in the satellite imagery, machine vision, fingerprints and facial recognition, agricultural photography images and many other applications.

2.4.2

Image Thresholding The threshold is one of the methods that are widely used for Image segmentation. It is

beneficial to distinguish foreground from the background. By identifying the minimum sufficient threshold (T) value and it can convert the image to a binary gray level image. It should be binary image to contain all the necessary information about the position and shape of units of interest. First obtaining the binary image feature is that it reduces the complexity of data and simplifies the process of recognition and classification[10]. Image thresholding is an image segmentation method that works with gray level images. The idea is to find a threshold and if the pixel is below the threshold value, it is considered as a background, otherwise it is considered as part of an object. For instant, the image in Figure 2.3(a) has one object and background. The result of image segmentation is shown in Figure 2.3(b). [11]

12

Figure 2.3: Original image (a) and result of segmentation (b). [11]

In addition, the threshold algorithms can be divided into two categories which are single level threshold and multi-level threshold. The approach of multi thresholding can generalized the thresholding of the image by finding out the multiple thresholds that’s aiming for separating the multi objects. Therefore, when an image has objects and background the threshold is used to segment that image. Moreover, the image processing is very significant in medical use. So in order to reconstruct 3D model for cancellous bone or tomography image we can use the image techniques such as micro-CT which always been as the proper techniques to be used. This technique is introduced also to discover 3D structural bone design [12]. The most common way to convert a gray-level image to a binary image is to select a single threshold value (T). Then all the gray level values below this T will be classified as black (0), and those above T will be white (1). The segmentation problem becomes one of selecting the proper value for the threshold T. [10]

13

In addition, threshold technique is one of the important techniques in image segmentation. This technique can be expressed as: [10] 2.4

Where: T is the threshold value. x, y are the coordinates of the threshold value point. P(x, y), f(x, y) are points the gray level image pixels. Threshold image g (x, y) can be defined:[10]

2.5

Therefore, the threshold algorithms can be divided into two categories which are single level threshold and multi-level threshold

2.4.3 Single Thresholding In single threshold the greyscale image is converted to binary which means black and white image that’s by selecting a grey level T from the original images and converting every pixel white or black regarding to whether is grey value is less or greater than T. It will be “white” if grayscale > T And “black” if its gray level is