CHAPTER 1 INTRODUCTION 1

CHAPTER 1
INTRODUCTION
1.1 INTRODUCTION
Brain tumors do not discriminate. Beginning in the brain they tend to stay there in people of all age groups but are statistically more frequent in older adults and children. Metastatic brain tumors i.e. which begin as cancer else where in the body and reach to the brain are more common in adults.

As per the American Brain Tumor Association 80,000 primary brain tumor cases are predicted to be diagnosed this year in America alone.

Gliomas are the most common type of tumor and surgery is the most common type of treatment but radiations and chemotherapy may also be used to slow down the growth of the tumors that cannot be physically removed. Magnetic Resonance images imaging may be used to provide detailed images of brain.

Tumors such as meningiomas, gliomas and glioblastomas can be easily segmented using the method of image processing proposed in here. Tumors can be anywhere in the brain in any shape and size. Gray scale values of MRI may vary depending on field of view, voxel resolution, gradient strength, type of MRI machine (1.5 Tesla, 3 Tesla, 7 Tesla) which varies for different hospitals.

Goal of our project is to identify a tumor from a given MRI scan of the brain with its location and extension using digital image processing which is done by identifying abnormal high intensity areas when compared to normal tissues.

1.2 Magnetic Resonance Imaging
Magnetic Resonance Imaging is based on the theory that protons and neutrons of the nucleus have an angular momentum which is called spin. When the number of subatomic particles is even the spins are cancelled whereas nuclei with odd number of subatomic particles will have a spin.

The MRI scanner uses powerful magnets to polarize and excite hydrogen proton in human tissue which produces a signal that can be encoded and detected, resulting in the images of the body being targeted. The part of the body that is to be examined is subjected to radio frequencies emitted by MRI machine which binds only to the hydrogen. Due to the radio frequency pulse, protons in the targeted area absorb the energy needed to make them spin in a specified direction.

MRI uses three electromagnetic fields which is static field (strong static electromagnetic field which polarizes the hydrogen nuclei), gradient field (weaker and time varying field) and weak radio frequency field used for manipulation of hydrogen nuclei to produce measurable signals which are received via radio frequency antennae.

1.3 Challenges
Brain is the key part of the central nervous system. Abnormal and uncontrolled cell division in the brain leads to brain tumor. We have used axial view of the brain from the MRI scan as MRI scan is less harmful as compared to CT scan. Several techniques such as imaging, biopsy, MRI, CT scan of the brain can be done based on the symptoms described by the patient. In biopsy a specimen of the brain tissue which is considered to be an tumor is taken by the pathologist and looks under the microscope to check for the presence of abnormalities in the tissue cells. Though doctors will only be able to know that the patient has tumor or not using the biopsy. To get the exact location of the tumor for being operated, MRI images are taken.

Traditional methods in hospitals is to segment the image manually and this depends on how well the doctor can perceive the image to get the required region extracted out which is made difficult with the minute variations and resemblances and original parts in the image.

The shortage of radiologists and large number of MRIs to be analyzed makes the task labor intensive and also increases the cost. It also depends on the experience and expertise of the doctor examining the images. Estimates indicate that between 10% and 30% are missed by radiologists during the screening process.

1.4 RELATED WORK
Various approaches have been carried out in the field of brain tumor segmentation. Sindhushree. K.S, et. al1 have developed a brain tumor segmentation method on 2D MRI data. Also detected tumors are represented in 3 dimensional view.

High pass filtering, histogram equalization, thresholding, morphological operators and segmentation using connected component labeling was carried out to detect tumor. The 2-dimensional extracted data was reconstructed into 3-Dimensional volumetric data and the volume of the tumor was also calculated.

M.C. Jobin Christ and R.M.S. Parvathi 2 proposed a methodology that integrates K Means clustering with marker controlled watershed segmentation algorithm. The proposed approach is a two stage process. First K Means clustering is used to get a primary segmentation of the input image and secondly marker controlled watershed algorithm is applied to the primary segmentation to get the final segmentation.

P.Vasuda, S.Satheesh 3, proposed a technique to detect tumors from MRI images using fuzzy clustering technique. This algorithm uses fuzzy C-means but the major drawback of this algorithm is the computational time required. Classifiers are also known as supervised methods since they require training data that are manually segmented and then used as references for automatically segmenting new data. The use of the same training data for classifying a large number of images, may lead to biased result. Supervised segmentation method requires considerable amount of training and testing data which comparatively complicates the process.

Vinay Parameshwarappa and Nandish S. et al, 2014 in his paper “Segmented morphological approach to detect tumor in brain images” 4, they proposed an algorithm for segmented morphological approach.

Sentilkumaran N and Thimmiaraja et al, 2014, Compare the image enhancement techniques in his paper “Histogram equalization for image enhancement using MRI brain images” 5, they presented the study of image
enhancement techniques and comparison of histogram equalization basic method like Brightness preserving adaptive histogram equalization (AHE), Local histogram equalization (LHE), global histogram equalization (GHE), Dynamic histogram equalization using different quality objective measures in MRI images.

CHAPTER 2
PROPOSED METHODOLOGY
PROPOSED METHODOLOGY
The part of the image the tumor normally has more intensity then the other portion and we can assume the area, shape and radius of the tumor in the image. We have used these basic conditions to detect tumor in the selected MRI images and the code goes through the following steps:
Preprocessing
Gray Scale Image
High Pass Filter
Enhanced Image
Post Processing
Threshold Segmentation
Watershed Segmentation
Morphological Operators

The proposed block diagram is shown below:
2762250118406
Figure 1: Proposed Block Diagram
PREPROCESSING
In preprocessing some basic image enhancement and noise reduction techniques are implemented. Apart from that different ways to detect edges and doing segmentations have also been used. The purpose of these steps is to improve the image and the image quality to get more surety and ease in detecting the tumor. The basic steps in preprocessing are the following:-
Image is converted to gray scale image in first step.

Noise is removed if any
The obtained image is then passed through a high pass filter to detect edges
Then the obtained image is added to original image to enhance it.

222631043878500The images below show the different preprocessing steps:
123190045339000416623546926500
Figure 2: MRI Images after being converted to gray scale, passing from high pass filter and finally the enhanced image.

2.3 POST PROCESSING
It is the phase of the implementation where main tumor detection will take place. It consist of three steps namely, threshold segmentation, watershed
segmentation and morphological operator.

Threshold Segmentation
Image thresholding is a simple yet effective way of partionioning an image into a foreground and background. This image analysis technique is a type of image segmentation that isolates objects by converting grayscale images into binary images. It is most effective in images with high levels of contrast. Thus in the MRI image segmentation is done on the basis of a threshold due to which whole image is converted into a binary image.

The images below shows the original image (Figure 3) and the image after applying threshold segmentation (Figure 4) using the code snippet:
T = graythresh(c); bw = im2bw(c,T+0.3); imshow(bw);
Figure 3: Original Image

Figure 4: After applying threshold segmentation
2.3.2 Watershed segmentation
The term watershed refers to a ridge that divides areas drained by different river systems. The catchment basin is the geographical area draining into a river or reservoir.

213868046863000
Figure 5: Image displaying watershed line and catchment line (Courtesy: mathworks.com)
Watershed Segmentation is the best method to segment an image or to separate a tumor but it suffers from over and under segmentation as every region makes its own catchment area, due to which we have used it as a check to our output. We have not used watershed segmentation on our input, rather it is only used on our output to check if the result is correct or not.

.2494279152442
(Figure 6: Displaying Watershed Segmentation)

2.3.2 MORPHOLOGICAL OPERATOR
Morphology is the study of shapes and structures from a scientific perspective. Morphological filters are formed from the basic morphology operations. A structuring element is mainly required for any morphological operation. Morphological operations operate on two images, structuring element and the input image. Structuring elements are small images that are used to probe an input image for properties of interest. Origin of a structuring element is defined by the centre pixel of the structuring element. In morphology, the structuring element defined will pass over a section of the input image where this section is defined by the neighbourhood window of the structuring element and the structuring element either fits or not fits the input image. Wherever the fit takes place, corresponding image that represents the input image’s structure is got suppression of the geometric features of the input
image that doesn’t fit the structuring element’s neighbourhood takes place. Two main morphology operations are erosion and dilation where erosion results in the thinning of the objects in the image considered and dilation results in thickening of the objects in the image. Dilation uses the highest value of all the pixels in the neighbourhood of the input image defined by the structuring element and erosion uses the lowest value of all the pixels in the neighbourhood of the input image. The basic purpose of the operations is to show only that part of the image which has the tumor that is the part of the image having more intensity.

The basic commands used in this step are:
Strel
Imerode
Imdilate
For a better understanding the code snippets are given below with their outputs.

SE = strel(‘disk’,0);
bw1 = imerode(bw,SE); imshow(bw1);
SE = strel(‘disk’,0); bw1 = imdilate(bw1,SE); imshow(bw1);

Figure 7: Applying snippet (A) Figure 8: Applying Snippet (B)
Figure 9 : (Imerode With strel(‘disk’,6);)

Figure 10: (Imdilate With strel(‘disk’,6);)
CHAPTER 3
SOFTWARE REQUIREMENT SPECIFICATION
3.1 PURPOSE
The purpose of this project is to create an application that allows the user to select the MRI image with the help of the user friendly interface and get the output which shows the tumor and its area if present in the MRI.

3.2 SCOPE OF THE PROJECT
The application developed in this project has the potential to be used as a means to analyze MRI images for tumor after performing rigorous testing.

3.3 TECHNOLOGY USED
Matlab 7.12.0 is used for developing the application and the user friendly graphical user interface. MATLAB is a high-performance language for technical computing. It integrates computation, visualization, and programming in an easy-to-use environment where problems and solutions are expressed in familiar mathematical notation. MATLAB is an interactive system whose basic data element is an array that does not require dimensioning. This allows you to solve many technical computing problems, especially those with matrix and vector formulations.

3.4 SYSTEM REQUIREMENT
The application was developed using Matlab 2011. Any device that can support Matlab 2011 or its superior versions along with necessary memory space and disk storage space will be able to execute our application.

DESCRIPTION OF MAIN FUNCTIONALITY
This application allows user to select a MRI image and by clicking on the output button the interface displays the tumor region in the MRI if present.

The output is displayed within seconds with preprocessing and post processing taking place in backend.

Figure 11: Graphical User Interface for the application
CHAPTER 4
SCREENSHOTS ; OUTPUT
4.1 The Console
100965044704000
Figure 12: The console window for the application
114490545466000Loading/Selecting the MRI image
Figure 13: The MRI image is loaded into the console detect tumor
Figure 14: Clicking on detect tumor generates the images for different functions
Figure 15: The tumor is shown in the output image.

Other Results: Different types of MRI images taken different angles are passed through this application and the tumor region is successfully found as shown in the results below.

Figure 16: Colored MRI image and its output

Figure 17: Different location and size of tumor
Figure 18: Different location and size of tumor

Figure 19: Different location and size of tumor
Figure 17,18,19 and 20 clearly shows that the application is able to find the tumors located at different places of brain and of different sizes.

Below are few mapped resultant tumor image onto the original grayscale image
20
lab code is able to distinguish between edema and tumor and does not return edema as tumor.

For all the MRI images the output has always been for correct irrespective of the size and location of the tumor.

CONCLUSION
Although there are many applications that implement brain tumor segmentation but this concept helps us to overcome many shortcomings of the popular methods such as cost and time as the methods pertaining to convolution neural network, fuzzy logics etc requires lot of complex operations which turns out to be costly and time consuming.

To improve the productivity and accuracy image processing can be combined with other soft computing techniques to provide a better and more accurate results with numerical values.

5.3 FUTURE WORK
With more research and rigorous testing this project can be used in future to identify brain tumors in low and high grade of MRI images. The code can also be embodied in MRI machines which also displays the result for tumor presence.

REFERENCES
Sindhushree. K. S, Mrs. Manjula. T. R, K. Ramesha, Detection And 3d Reconstruction Of Brain Tumor From Brain Mri Images, International Journal of Engineering Research ; Technology (IJERT), vol. 2, no. 8, pp 528-534, 2013.

M.C. Jobin Christ, R.M.S.Parvathi, “Segmentation of Medical Image using Clustering and Watershed Algorithms”, American Journal of Applied Sciences, vol. 8, pp 1349-1352, 2011.

P.Vasuda, S.Satheesh, “Improved Fuzzy C-Means Algorithm for MR Brain Image Segmentation,International Journal on Computer Science and Engineering (IJCSE), vol. 02, no.05, pp 1713-1715, 2010.

Vinay Parmeshwarappa, Nandish S, “A segmented morphological
approach to detect tumor in brain images”, IJARCSSE, ISSN: 2277 128X , volume 4, issue 1, January 2014.

Senthilkumaran N, Thimmiaraja J,”Histogram equalization for image enhancement using MRI brain images”, IEEE CPS,WCCCT.2014.45.

Pratibha Sharma, Manoj Diwakar, Sangam Choudhary, “Application of Edge Detection for Brain Tumor Detection”, International Journal of Computer Applications, vol.58, no.16, pp 21-25, 2012.

Anam Mustaqeem, Ali Javed, Tehseen Fatima, “An Efficient Brain Tumor Detection Algorithm Using Watershed ; Thresholding Based Segmentation”, I.J. Image, Graphics and Signal Processing, vol. 10,no. 5, pp 34-39, 2012.

Chang Wen Chen, Jiebo Luo, Kevin J. Parker,”Image Segmentation via Adaptive K-Mean Clustering and Knowledge-Based Morphological Operations with BiomedicalApplications”, IEEE Trans. Image Process.,vol.7,no.12, pp1673-1683, 1998.

P.Dhanalakshmi , T.Kanimozhi, “Automatic Segmentation of Brain Tumor using K-Means Clustering and its Area Calculation”, International Journal of Advanced Electrical and Electronics Engineering ,vol. 2,no. 2,pp 130- 134, 2013.

G.Evelin Sujji, Y.V.S Lakshmi, G.Wiselin Jiji,”MRI Brain Image Segmentation based on Thresholding”, International Journal of Advanced Computer Research, vol.3 no.1, pp 97-101, 2013.

Swe zin Oo, Aung Soe Khaing, “Brain tumor detection and segmentation using watershed segmentation and morphological operation”, IJRET, Volume 03,Issue 03, March 2014.

Rajesh Kumar Rai, Trimbak R. Sontakke, “Implementation of Image Denoising using Thresholding Techniques”, International Journal of Computer Technology and Electronics Engineering (IJCTEE),vol.1,no. 2, pp 6-10, 2013.

Mohammed Y. Kamil, “Brain Tumor Area Calculation in CT-scan image using Morphological Operations”, IOSR Journal of Computer Engineering (IOSR-JCE) ISSN: 2278-8727, Volume 17, Issue 2, Ver.-V, pp 125-128, Mar
– Apr. 2015.

Poonam, Jyotika Pruthi, “Review of image processing techniques for automatic detection of tumor in human brain”, IJCSMC, Vol.2, Issue.11, pp.117-122, November 2013.

A.Sindhu1 , S.Meera2 “A Survey on Detecting Brain Tumor in mri Images Using Image Processing Techniques “International Journal of Innovative Research in Computer and Communication Engineering Vol. 3, Issue 1, January 2015.

Team Details:
Project Member name : signature:

Utsav rawat (1501051147)
Manish bhati (160105900)
Kundan singh (1501051198)
Rashesh Dwivedi (1501051158)
Guide Name and Signature:
Mr. Nishant Singh Rathore
Assistant Professor
IT Department
DIT University

Signature(guide): ______________