DETECTION OF TUMOUR IN 3D BRAIN IMAGE USING MRI MINING TECHNIQUES

Abstract- Automated detection of tumours in medical image is motivated by the necessity of high accuracy when deals with human career. This paper deals with how the brain tumor is detected using Magnetic resonance imaging (MRI) and Computed Tomography (CT) Technology. The double reading of medical image could be lead to better tumour detection. K means algorithm handle large set of data by finding centroid. By using this algorithm , locating initial seed point is easy and which will give more accurate and high resolution result. K different tissue type which include gray, white matter and CSF and other abnormal tissue. MR images can be either scale or multivalued. Current applications of WM tractography include surgical planning. This paper proposed an algorithm which we use image processing technique to classify the images either as normal or abnormal then segment the tissues of abnormal brain MRI to identify brain tumor. Keywords: Image and data mining, MRI, CT , Segmentation, and Cerebrospinal fluid (CSF).

1. INTRODUCTION
The X-ray angiography (XRA), magnetic resonance angiography (MRA), magnetic resonance imaging (MRI), computed tomography (CT), and other imaging modalities are heavily used in clinical practice. Such images provide complementary information about the patient. While increased size and volume in medical images required the automation of the diagnosis process, the latest advances in computer technology and reduced costs have made it

possible to develop such systems. Blood vessel delineation on medical images forms an essential step in solving several practical applications such as diagnosis of the vessels (e.g. stenosis or malformations) and registration of patient images obtained at different times. Segmentation algorithms form the essence of medical image applications such as radiological diagnostic systems, multimodal image registration, creating anatomical atlases, visualization, and computeraided surgery. Radiology department are at the center of a massive change in technology. The ubiquitous radiographic film that has been the basis of image management for almost 200 year is being displaced by new digital imaging modalities such as 1.Computed tomography 2.Magnetic resonance 3. Nuclear medicine 4.Ultra sound 5. Digital tomography 6.Computed tomography 7.MR Spectography. These digital modalities are continuously refined and new digital applications are being developed. Among this medical imaging techniques MRI , x-ray computer tomography, positron emission tomography(PET) and ultrasound has given physician a non invasive means. Vessel segmentation algorithms are the key components of automated radiological diagnostic systems. Segmentation methods vary depending on the imaging modality, application domain, method being automatic or semi-automatic, and other specific factors. There is no single segmentation method that can extract vasculature from every medical image modality. While some methods employ pure intensity-based pattern recognition techniques such as thresholding followed by connected component analysis , some other methods apply explicit vessel models to extract the vessel contours. Depending on the image quality and the general image artifacts such as noise, some segmentation methods may require image preprocessing prior to the segmentation algorithm. On the other hand, some methods apply post-processing to overcome the problems

arising from over segmentation. vessel segmentation algorithms and techniques can be divided into six main categories, pattern recognition techniques, model-based approaches, tracking-based approaches, artificial intelligencebased approaches, neural network-based approaches, and miscellaneous tube-like object detection approaches. Clustering analysis plays an important role in scientific research and commercial application. This thesis provides a survey of current vessel segmentation methods using clustering approach and provides both early and recent literature related to vessel segmentation algorithms and techniques.

determination of point to point mapping that alings the anatomy in one image with that in another.

2. BACKGROUND
Medical images are very important in medical diagnosis and treatment. These images are different from typical photographic images. Numerous techniques developed and reported in the literature for segmenting MR brain images. Most of them are based on clustering voxels with similar features as the same tissue type. Methods that rely exclusively on partitioning the feature space such as the histogram or scatter gram work well only for MR images with excellent soft tissue contrast, high signal-to noise ratio (SNR) and negligible shading effect. Unfortunately, most of clinically available MR images do not yet meet such a quality standard. It has been shown that the incorporation of spatial interactions between neighboring voxels into the segmentation procedure reduces significantly the degradation of segmentation outcome due to poor SNR and feature contrast. In contrast, edge based segmentation techniques, which explore local intensity contrast to partition the image into regions through the Marr-Hildreth operator or graph theory, are insensitive to the shading effect. Image segmentation is the process of automatic extraction of structures of interest from the images. Image registration is the process of

Fig.2.1. Hierarchical approach evaluation. (a) A pair of model (blue)-target (green) brains. (b)(d) Ground truth (blue), segmented fibers obtained with hierarchical ICF (red), without hierarchical ICF (yellow) in axial view (b), coronal view (c), and sagittal view (d). 2.1 MRI System A MRI technology have following components 1. a large magnet to generate the magnetic field,2. shim coils to make the magnetic field as homogeneous as possible,3.a radio frequency (RF) coil to transmit a radio signal into the body part being imaged, 4.a computer to reconstruct the final image. 2.3 MRI Planes: MRI images are divided into three planes: sagittal plane, horizontal plane, and coronal plane. Mining medical images involves many process The process to be depend upon type and complexity of the imageto be mined. It is simpler to mined 2D x-ray as compared to 3D CT scan of the images.

3. MRI MINING

Data Base

User login

Segmen tation of 3D Brain Using Threshol d Value

View User Details

next step is to take each point belonging to a given data set and associate it to the nearest centroid. When no point is pending, the first step is completed and an early groupage is done. At this point we need to re-calculate k new centroids as barycenters of the clusters resulting from the previous step. After we have these k new centroids, a new binding has to be done between the same data set points and the nearest new centroid. A loop has been generated. As a result of this loop we may notice that the k centroids change their location step by step until no more changes are done. In other words centroids do not move any more. Finally, this algorithm aims at minimizing an objective function, in this case a squared error function.

Fig 3.1 User Details and Report 3.1 Preprocessing Pre-process technology enhances class 1 semantic content immediately before images are sent to an 3D Segmentation Process. Semantic content in a class 1 sense refers to the correct classification of a blob of pixels as a specific character (or part of such), a line, a part of an image, or noise. By implementing several powerful and proprietary noise removal and Pixel grouping enhancement algorithms 3.3 Segmentation Segmentation refers to the process of partitioning a digital image into multiple regions (sets of pixels). The goal of segmentation is to simplify and/or change the representation of an image into something that is more meaningful and easier to analyze. Image segmentation is typically used to locate objects and boundaries (lines, curves, etc.) in images.The result of image segmentation is a set of regions that collectively cover the entire image, or a set of contours extracted from the image (see edge detection). Each of the pixels in a region are similar with respect to some characteristic or computed property, such as color, intensity, or texture. Adjacent regions are significantly different with respect to the same characteristic(s).

3.2 Centroid Calculations K-means is one of the simplest unsupervised learning algorithms that solve the well known clustering problem. The procedure follows a simple and easy way to classify a given data set through a certain number of clusters (assume k clusters) fixed a priori. The main idea is to define k centroids, one for each cluster. These centroids shoud be placed in a cunning way because of different location causes different result. So, the better choice is to place them as much as possible far away from each other. The

4. K-MEANS ALGORITHM
The improved K-means algorithm is a solution to handle large scale data, which can select initial clustering center purposefully, reduce the sensitivity to isolated point, avoid dissevering big

cluster. By using this technique locating the initial seed point is easy and which will give more accurate and high-resolution result.By using various techniques we can study or compare the results and find out which technique gives higher resolution Initial centroid algorithm is useful to avoid the formation of empty clusters, as the centroid values are taken with respect to the intensity value of the image. Proposed algorithm is better for large datasets and to find initial centroid. Initial centroid algorithm is useful to avoid the formation of empty clusters, as the centroid values are taken with respect to the intensity value of the image. Proposed algorithm is better for large datasets and to find initial centroid. Decide the class membership of the N objects by assigning them to the nearest cluster center. Re-estimate the K cluster center by assuming the memberships found above are correct

The objective of this paper is to propose a new image mining process model to classify normal or abnormal images and develop efficient algorithms to achieve high quality segmentation of WM ,GM ,and CSF from MR brain images /volumes which in turn can be used to diagonise brain related diseases. This study significant for the following reasons: The proposed algorithm can assist physicians improve the accuracy and speed of decision making based on medical images in medicine, thereby leading to dramatic improvement in health care. Experiences gained in applying he proposed algorithm to medical image data can provide the catalys to promote knowledge extraction from image data sets in other domains.

6. SYSTEM OVERVIEW
The advantage of MRI are the spatial resolution is high and provide detailed image. MRI are used in detecting brain tumour and tracking them. The tracking of the tumour is important especially patient is under treatment in order to observe the changes. The denoising method are used in : Wavelet based method and Histogram equalization method.

K-MEANS ALGORITHM: 1. Decide on a value of K, the number of clusters. 2. Initialize the K cluster centers (randomly if necessary). 3. Decide the class membership of the N objects by assigning them to the nearest cluster center. 4. Re-estimate the K cluster center by assuming the memberships found above are correct. 5. Repeat 3 and 4 until none of the N objects changed membership in the last iteration.

7. CONCLUSION
Vessel segmentation methods have been a heavily researched area in recent years. Even though many promising techniques and algorithms have been developed, it is still an open area for more research. This algorithm does not require any user interaction, not even to identify a start point. Here seed points are selected randomly which determines the main branches of the vessel structure. Random selection of seed points does not yield accurate segmentation. Accuracy of the segmentation process is essential to achieve more precise and repeatable radiological diagnostic systems. Accuracy can be improved by incorporating a priori information on vessel anatomy and let high level knowledge guide the

Table 4.1 k-means algorithm:

5. RESEARCH OBJECTIVES

segmentation algorithm. k-means algorithm is a popular clustering algorithm applied widely, but the standard algorithm which selects k objects randomly from population as initial centrsoids can not always give a good and stable clustering. Experimental results show that selecting centroids by our algorithm can lead to a better clustering. Along with the fast development of database and network, the data scale clustering tasks involved in which becomes more and more large. K-means algorithm is a popular partition algorithm in cluster analysis, which has some limitations when there are some restrictions in computing resources and time, especially for huge size dataset. The improved K-means algorithm presented in this paper is a solution to handle large scale data, which can select initial clustering center purposefully, reduce the sensitivity to isolated point, avoid dissevering big cluster, and overcome deflexion of data in some degree that caused by the disproportion in data partitioning owing to adoption of multi-sampling.Our ongoing research focuses on the development of methods to segment coronary arteries in a sequence of angiographic images while preserving the topology of the vessel structure.

[2].Andy Harris, MICROSOFT C# PROGRAMMING, Prentice hall of India Pvt ltd., [3].R.De La Paz, E. Herskovits, V.Gesu, et al. Cluster analysis of medical magnetic resonance images MRI Data: Diagonostic application and evaluation, SPIE Extracting meaning from Complex Data: Processing, Display, Interaction,Vol.1259,1990,pp.176-181. [4].V. Megalooikonomou, J. For, L. Shen, F Makedon. Data Mining in Brain imaging. Statistical methods in medical research 2000, [5].Roger S.Pressman, SOFTWARE ENGINEERING, TataMcGraw Hill Publications, fifth edition. [6].k. LeeIntelligent Shape based Image Analysis , Department of computer science ,University of Iowa , 55 :247 Project 2001. [7].Elias M.Awards, SYSTEM ANALYSIS AND DESIGN, Galgotia Publications Private Limited Companies, 1997 Edition. [8].Gennekkan. Computer Aided Diagonisis in Chest Radiography. PhD thesis, Utrecht University, march,2001 pp.14-17. [9].Herbert Schildt, THE COMPLETE REFERNCE C# 2.0, TataMcGraw Hill Publications, second edition. [10].N. Lavac, E. Keraunou and B. Zupan. Intelliogent Data Analysis in Medicine and Phrmacology. (IDAM4P-97),Nagoya, Japonska, Kluwer Acedamic Publishers,1997,pp 61-67. [11].V.K.Jain, THE COMPLETE GUIDE TO C# PROGRAMMING, Dreamtech press.

8. FUTURE ENHANCEMENT
The improved K-means algorithm is a solution to handle large scale data, which can select initial clustering center purposefully, reduce the sensitivity to isolated point, avoid dissevering big cluster. By using this technique locating the initial seed point is easy and which will give more accurate and high-resolution result.By using various techniques we can study or compare the results and find out which technique gives higher resolution

9. REFERENCES
[1]. Simon Robinson, Christian Nagel, Karli Watson, PROFESSIONAL C#, Wiley Dreamtech India Pvt ltd., third edition.

[12].Rui J. P. Defigueiredo, w.Rodman Shanklet, Andrea Maccato,Malcolm b. Dickt, Prashanth Mundkur, Neural_network_based cclassification of cognitively normal, demented, Alzheimer disease and vasculear dementia from single photon emission with computed tomography image data from the brain.

Pattern Anal. Mach. Intell., vol. 14, no. 2, pp. 239256, 1992. [16] I. Shimizu, Z. Zhang, S. Akamatsu, and K. Deguchi, Headpose determination from one image using a generic model,in Proc. IEEE Int. Conf. Automatic Face Gesture Recognit., 1998, pp. 100105. [17] G. Stylianou, A feature based method for rigid registration of anatomical surfaces, in Geometric Modeling for Scientific Visualization, G. Brunnett, B. Hamann, H. Mller, and L. Linsen, Eds. New York: Springer, 2003, pp. 139152. [18] J. Shimony, A. Snyder, N. Lori, and T. Conturo, Automated fuzzy clustering of neuronal pathways in diffusion tensor tracking, in Proc. Int. Soc. Magn. Reson. Med., Honolulu, HI, 2002, pp. 10.

[13] M. Catani, R. J. Howard, S. Pajevic, and D. K. Jones, Virtual in vivo interactive dissection of white matter fascicule in the manbrain, Neuroimage, vol. 17, no. 1, pp. 7794, 2002.
[14]. A. Mayer and H. Greenspan, Bundles of interest based registration of white matter

tractographies, in Proc. IEEE Int. Symp. Biomed. Imag.: From Nano to Macro., 2008, pp. 919922. [15] P. J. Besl and N. D. Mckay, A method for registration of 3-D shapes, IEEE Trans.