Published on

Published in: Technology, Health & Medicine
  • Be the first to comment

  • Be the first to like this


  1. 1. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 - 6375(Online), Volume 5, Issue 3, March (2014), pp. 122-131 © IAEME 122 BRAIN TUMOR AND EDEMA DETECTION USING MATLAB Nidhi1 , Poonam Kumari2 1, 2 UCIM/CIL/SAIF, Panjab University, Chandigarh ABSTRACT Advanced techniques of medical image processing and analysis find widespread use in medicine. Various imaging modalities like CT scan, MRI, ultrasound are being used for imaging brain tumors. In recent years, MRI has emerged as the best for clear identification of cancer and other anomalies in breast, prostate, liver, brain etc. The tumor detection becomes more complicated for the huge image database especially when edema is present with the tumor. So a software approach is needed to aid the accurate, faster clinical diagnosis. Proposed work focuses on the detection of brain tumor and edema from MRI images using MATLAB and clear distinction between tumors and edema. The objective is to provide advanced image processing tools in a format that is user friendly and is inexpensive too. The study aims to introduce an algorithm which incorporates useful operations on the MRI brain image including filtering, enhancement, arithmetic operations, segmentation, extracting region of interest, and morphological operations. Here we detect the tumor and edema, segment them and the final image having clear boundary between edema and tumor is superimposed on the original image to highlight the tumor and edema boundaries. Our study aims to help the physician for surgical planning. KEYWORDS: Edema, MRI images, co-resemblance, Histogram Equalization, Segmentation, Sobel Edge Detection Filter, Image Superimposition. 1. INTRODUCTION Brain is the central processing unit of world’s most complicated machinery, that is, human being. Brain acts as the in charge of human thoughts, feelings, speech, and memory and also plays a pivotal role in controlling muscle movements. Brain helps in the interpretation of sensory information. A tumor is an abnormal new mass of tissue that serves no purpose. The term brain tumor is used to describe any tumor growing within the skull, though a more accurate term might be intercranial tumor. Brain tumor is defined as any intercranial tumor created by abnormal and uncontrolled cell division, normally either in the brain itself(neurons, glial cells, lymphatic blood INTERNATIONAL JOURNAL OF COMPUTER ENGINEERING & TECHNOLOGY (IJCET) ISSN 0976 – 6367(Print) ISSN 0976 – 6375(Online) Volume 5, Issue 3, March (2014), pp. 122-131 © IAEME: www.iaeme.com/ijcet.asp Journal Impact Factor (2014): 8.5328 (Calculated by GISI) www.jifactor.com IJCET © I A E M E
  2. 2. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 - 6375(Online), Volume 5, Issue 3, March (2014), pp. 122-131 © IAEME 123 vessels), in the cranial nerves(myelin), in the brain envelopes(meninges), skull, pituitary and pineal gland or spread from cancers primarily located in other organs [1]. The symptoms of brain tumor depends on tumor size, type and location. Some common symptoms of brain tumor are- • Headaches. • Nausea and vomiting. • Changes in speech, vision or hearing. • Problems in walking. • Seizures or convulsions. • Changes in mood, personality or ability to concentrate. • Problems with memory. 1.1 Common types of brain tumor A brain tumor may be of primary or secondary type depending on its location of origin. Primary tumors originates in the brain itself while the secondary tumors originates in some other part of body and then spread to brain. There are two categories of brain tumors according to the most commonly used classification- Benign- Benign tumors are non-cancerous mass of cells that grows slowly in the brain. It usually stays in one place and does not spread. These tumors can be removed and they seldom grow back. Most of the benign brain tumors are detected by Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) scans. Benign tumors, however, can be life threatening because they can compress brain tissues and other structures inside the skull. The following are the most frequently diagnosed benign brain tumors- Meningioma, Schwannoma, Pituitary adenomas, Hemangioblastomasa, Craniopharyngioma [2][3]. Malignant- A malignant brain tumor is a rapidly growing cancer that spreads to other areas of the brain and spine. Most of the malignant brain tumors are secondary but can be primary too. These tumors are life threatening. Common malignant brain tumors are- Gliomas, Ependymomas, Oligodendrogliomas, Mixed gliomas [2][3]. 1.2 Edema Edema is commonly known as brain swelling which can occur in specific location in vicinity of the brain tumor or throughout the brain. It is the “extra fluid” within the tissue of the brain. Edema increases intercranial pressure which can prevent blood from flowing to the brain, thus depriving it of the oxygen it needs to function. Damage or death of brain cells may result [4]. 1.3 Diagnosis of brain tumor and edema One or more of the following methods may be used to detect the presence of a brain tumor having edema and if it has spread- • Biopsy. • Sterotactic Biopsy. • Surgery. • Lumbar Puncture. Imaging methods • Computed Tomography (CT) scan. • Magnetic Resonance Imaging (MRI). • Positron Emission Tomography (PET) scans. • Diffusion Tensor Imaging (DTI).
  3. 3. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 - 6375(Online), Volume 5, Issue 3, March (2014), pp. 122-131 © IAEME 124 The MRI method is the best in detecting brain tumors due to its high resolution and ability to show clear brain structures, tumor’s size and location. MRI makes use of the property of Nuclear Magnetic Resonance to image nuclei of atoms inside the body. It provides a good contrast between different soft tissues of the body which makes it useful in imaging brain cancers. MRI image of brain having tumor with edema is used in the proposed work [1]. 3. PROBLEM FORMULATION Detection of brain tumor having edema as its most prominent feature is a serious issue in imaging science. Generally, the grading and analysis of brain tumors having edema is done by doctors by simply viewing the image scans which is a very difficult task due to very minute variations. Based on the expert’s analysis and detection, the tumor excision in surgery is performed and it can have chances to get the false detection. The problem of manual detection becomes more severe when the database is too large. An important step in analysis of MRI brain images is to extract the boundary of the tumor part which becomes more complicated when tumor have edema in its vicinity. Due to co-resemblance between tumor part and edema, the biological analysis thus becomes prediction of affects. To solve the problem, the proposed work describes the strategy for detection, segmentation and feature extraction of brain tumor part and edema in an easy to use, inexpensive format using MATLAB software. This software based approach aims to introduce an algorithm for detecting and segmenting the brain tumor and edema from normal brain using basic image processing operations( preprocessing, enhancement, segmentation, morphological operations, feature extraction) in MATLAB 4. METHODOLOGY The proposed algorithm follows the following sequence of steps- Image Acquisition Data Import Convert original image to grayscale image. Preprocessing Image Enhancement Image Segmentation Morphological Operations Final image is superimposed on original Image Brain tumor and edema detected
  4. 4. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 - 6375(Online), Volume 5, Issue 3, March (2014), pp. 122-131 © IAEME 125 4.1 Image Acquisition- Image acquisition in image processing can be broadly defined as the action of retrieving an image from source, usually a hardware based source, so it can be passed through whatever processes need to occur afterward[5]. Image acquisition is the first step in the proposed workflow sequence. MRI brain images are used in the proposed system. MRI uses magnetic field and radio waves to provide detailed information about brain tumor with edema anatomy, cellular structure and vascular supply, making it an important tool for the effective diagnosis, treatment and monitoring of the brain diseases [6]. The images used in the present study are acquired from Satyakiran Hospital, Sonepat, Haryana. 4.2 Importing data in MATLAB- Image processing toolbox in MATLAB supports images generated by a wide range of devices, including, digital cameras, satellites, airborne sensors, medical imaging devices, microscopes, telescopes and other scientific instruments. Image processing toolbox shows compatibility with a number of specialized image file formats. For medical images, it supports DICOM files. There are several ways to import the data into MATLAB environment for processing. In the present study the database is in CD-ROM and stored on the desktop of computer, then the desktop path is entered in the address area of the MATLAB 4.3 Grayscale Imaging- Generally, grayscale imaging is sometimes called “black and white” but in technical terms it is a misnomer. The true black and white is known as halftone which consists of only possible shades of pure black and white. A grayscale image only consists of shades of gray without apparent color. When MRI images are viewed on computer screen, they look like black and white but in actual they contain some primary colors (RGB) content. So, for further processing of MRI brain image, it must be converted to perfect grayscale image in which the red, green and blue components all have equal intensity in RGB space(The lightness of the gray is equal to the number representing the brightness levels of primary colors. The brightness level of RGB is a number from decimal 0 t0 255 or binary 00000000 to 11111111. For every pixel in a RGB grayscale image, R=G=B where black is represented by R=G=B=0 and white by R=G=B=1)[6][7]. So it becomes easy to process grayscale images (only single intensity value is necessary for each pixel) as compared to full color image ( three intensity values are necessary, each for RGB, for a single pixel). The brain image received from MRI is converted to grayscale image by eliminating hue and saturation information while retaining the luminance. The original MRI brain image has properties 320x320x3 and conversion to grayscale image makes the properties 320x320. The grayscale brain image is then converted to double data class type. Fig.1: Original Image Fig.2: Grayscale Image
  5. 5. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 - 6375(Online), Volume 5, Issue 3, March (2014), pp. 122-131 © IAEME 126 4.4 Preprocessing- Preprocessing is the initial step for detecting brain tumor with edema. Basically, this process involves denoising the image and increasing the signal to noise ratio using different filtering techniques [8]. MRI images are bound to have some noise in them and white noise is one of the most common problems in processing MRI images. Noise may also be introduced due to motion artifacts (movement of patient during scan) in the MRI images. The most common methods used for preprocessing of medical images till now are the low pass filtering( for sharpening), averaging filters( for smoothening) , but these methods have certain drawbacks like LPF of MRI image may blur the edges while averaging filters may blur the details as well as edges in an image. Moreover, averaging filters are not as effective for impulse noise (salt and pepper noise) [5]. So to overcome above stated problems, the following preprocessing methods are used in the present study- 4.4.1 High Pass Filtering- It is done to sharpen the image and a high pass filter forms the basis for most of the sharpening methods. Sharpening filters are the most often and abused processing tools. When applied properly, sharpening can often improve apparent image quality by making it look crisper and more defined. However, not all sharpening methods are created equal. When performed too aggressively, unseen sharpening artifacts may appear. A high pass filter preserves the high frequency information within an image while reduce the low frequency information, thus emphasizing the transitions in the image intensities. It works by analyzing the values of each pixel in an image and changing it based on the values of its neighbors. In high pass filtering, the brightness of the centre pixel is increased relative to its neighboring pixels by the kernel of the filter. The kernel array consists of a single positive value at its centre, which is completely surrounded by negative values. 4.4.2 Median Filtering-It is done for smoothening of MRI brain image. Median filtering is very effective for removing “salt and pepper” noise (random occurrences of black and white pixels) [6]. It is somewhat like mean filter. However, it often does a better job than the mean filter by preserving useful detail in the image. The median filter considers each pixel in the image and looks at its nearby neighbors to decide whether or not it is representative of its surroundings [5][6][7]. The value of the output pixel is then determined by median of the neighborhood pixels. The median filter does not create new unrealistic pixel values when the filter straddles an edge in the image. For this reason, the median filter is much better at preserving sharp edges than other filters used till now ( like mean, average filters etc.). Fig.3 shows the impact of “salt and pepper” noise on the grayscale image and fig.4 depicts the noise free image of brain. Fig.3: Salt & Pepper noise in image Fig.4: Median filtered image
  6. 6. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 - 6375(Online), Volume 5, Issue 3, March (2014), pp. 122-131 © IAEME 127 4.5 Image Enhancement- Image enhancement brings out the details that are obscured and highlight certain features of interest in an image. The fundamental enhancement needed in the MRI images is the contrast enhancement. Contrast is the main reason for the co-resemblance between the tumor part and edema in the MRI brain image. Due to poor contrast, the tumor region and edema are considered the same by the doctors by manually viewing the MRI scans. Contrast between the brain, tumor part and edema may be present in the MRI image but below the threshold of human perception. Different methods have been proposed till now for image enhancement like basic gray level transformations, binarization etc. In the present approach two methods are used for enhancing the contrast of MRI brain image. 4.5.1 Arithmetic Operations-Arithmetic operations are performed on a pixel by pixel basis between two or more images. The actual mechanics of implementing arithmetic operations can be done sequentially, one pixel at a time, or in parallel, where all operations are performed simultaneously. There are total four arithmetic operations (addition, subtraction, multiplication and division) that can be applied on an image [5].In the present paper, subtraction operation is performed between the grayscale and the double class MRI brain image. The subtraction operation is expressed as- G(x,y) = F(x,y)-H(x,y) and the difference is obtained by computing the difference between all pairs of corresponding pixels from F and H. The key usefulness of subtraction is the enhancement of differences between images which is its main advantage over other methods used till now. 4.5.2 Histogram Equalization- Histogram equalization is a technique for adjusting image intensities to enhance contrast of an image. Better contrast is obtained via the histogram of the image, then using histogram equalization that allows the areas with low contrast to gain higher contrast by spreading out the most frequent intensity values. As it can be seen in the subtracted image that edema and tumor part have very close contrast values and histogram equalization then increases the global contrast of the MRI brain image. This method is used because the intensities can be better distributed on the histogram which represents the relative frequency of occurrences of various gray levels in the image. Histogram equalization is a three step process [8]- • Formation of histogram. • Calculation of new intensity values for each intensity level of image. • Replace the previous intensity values of the image with new calculated intensity values. The change in contrast between tumor part and edema is clearly shown in fig.5 and fig.6. The black colored region is edema and the grayish smoky part is the tumor region in brain.
  7. 7. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 - 6375(Online), Volume 5, Issue 3, March (2014), pp. 122-131 © IAEME 128 Fig.5: Subtraction image Fig.6: Histogram equalization 4.6 Image Segmentation- It is the division of an image into meaningful structures. Segmentation subdivides an image into its constituents regions or objects and it is an essential step in image analysis, object representation, visualization and many other image processing tasks [5]. More precisely, image segmentation is the process of assigning a label to every pixel in an image such that pixels with same label share certain identical visual characteristics. The resultant image is a set of segments that collectively cover the entire image or a set of contours extracted from the image. Generally, the doctor always need to have keen observation of the anatomical structure during the process of manual segmentation in MRI scans. But this process is too much time consuming and if the initial segmentation result is not correct than other subsequent results also produces incorrect measurement results. A great number of segmentation methods has been employed in the past decades for brain tumor segmentation like clustering methods, fuzzy logic approach, neuro-fuzzy approach, watershed segmentation, random walk etc. but these all methods produces unsatisfactory results due to unsharp edge boundaries and more time consumed to produce desired result [13]. Moreover, these methods can only segment the tumor region but edema present in the vicinity of tumor region can not be distinguished from the tumor part. In the present paper, we are using a filter to segment the tumor region and edema from the normal brain. The filter used is sobel edge detection filter which is commonly used in the computer vision. Sobel is a gradient operator which detects the edges by looking for the maxima and minima in the first derivative of the image and the result is a vector valued operator. Sobel operator applies gradient filter which average the image perpendicular in gradient direction. The main advantages of using sobel operator over other methods used till now are- • Errors in magnitude and angle are smaller than with discrete differences. • Smaller anisotropy. • Clear segment tumor part from edema which proves helpful in tumor excisions during surgery. Fig.7 and fig.8 are showing the boundaries between different segments in the MRI brain image.
  8. 8. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 - 6375(Online), Volume 5, Issue 3, March (2014), pp. 122-131 © IAEME 129 Fig.7: Sobel filter on subtracted image Fig.8: Sobel filter on histogram equalized image 4.7 Morphological operations- Morphological processing deals with the tools for extracting image components that are useful in the representation and description of shape. Basically, these operations are the linear operations related to the shape of features in an image. These operations process images based on shapes by applying a structuring element to the input image, creating an output image of the same size. In a morphological operation, the value of each pixel in the output image is based on a comparison of the corresponding pixel in the input image with its neighbors. Some operations test whether the element “fits” within the neighborhood, while others test whether it “hits” or intersects the neighborhood. By choosing the shape and size of neighborhood, a morphological operation can be constructed that is sensitive to specific shapes in the input image. The morphological operations used in the present paper are dilation (adds pixels to boundaries of objects in an image) and erosion (removes pixels from object boundaries). Dilation and erosion depends on size and shape of the structuring element used to process the image. Fig.9 shows the thickening of boundaries of segments in the image and fig.10 depicts the extraction of interested tumor and edema boundaries while removing all other unwanted boundaries. Fig.9: Dilated image Fig: 10 Unwanted boudaries removed 4.8 Image superimposition- The final segmented image is then superimposed on the original image which clearly distinguish between tumor and edema and the boundaries are detected which becomes more visible when superimposed on the anatomical structure of brain MRI image.
  9. 9. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 - 6375(Online), Volume 5, Issue 3, March (2014), pp. 122-131 © IAEME 130 Fig.11: Final superimposed image In this paper an algorithm in MATLAB has been developed to detect the brain tumor and edema from MRI brain images based on the sobel segmentation method. To summarize the developed method, the operations are performed on gray scale image of brain and then enhancement is done followed by segmentation and final superimposition of output image on original image. The tumor part is then separated from the edema in its vicinity by clear boundaries between the two which greatly aids the physicians in surgical procedures of tumor excisions. 5. FUTURE WORK The proposed system can be extended for some other imaging modality like CT, PET-CT, DTI etc, for different organs of human such as lungs, liver, breast and so on. This proposed work finds its wide applications in the Medical Imaging Sciences and other related research areas. REFERENCES [1] M.Karuna 1, Ankita Joshi 2. Automatic detection and severity analysis of brain tumors using GUI in MATLAB. IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163, pISSN: 2321-7308. [2] Kimmi Verma1, Aru Mehrotra2, Vijayeta Pandey3, Shardendu Singh4. Image processing techniques for the enhancement of brain tumor patterns. International Journal of Advanced Research in Electrical, Electronics and Instrumentation Engineering Vol. 2, Issue 4, April 2013. [3] Nagalkar V.J. and Asole S.S. Brain Tumor Detection Using Digital Image Processing Based On Soft Computing. Journal of Signal and Image Processing, ISSN: 0976-8882 & E-ISSN: 0976-8890, Volume 3, Issue 3, pp.-102-105. [4] Ashraf Anwar, Arsalan Iqbal . Image Processing Technique for Brain Abnormality Detection .International Journal of Image Processing (IJIP), Volume (7) : Issue (1) : 2013. [5] “Digital Image Processing”, 3/E by Rafael C. Gonzalez, Richard E. Woods, ISBN-10: 013168728X. [6] Rajesh C. Patil, Dr. A. S. Bhalchandra ,Brain Tumour Extraction from MRI Images Using MATLAB. International Journal of Electronics, Communication & Soft Computing Science and Engineering ISSN:2277-9477, Volume2, Issue1. [7] N.Gopinath, Extraction of Cancer Cells from MRI Prostate Image Using MATLAB. International Journal of Engineering Science and Innovative Technology (IJESIT) Volume 1, Issue 1, September 2012.
  10. 10. International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 - 6375(Online), Volume 5, Issue 3, March (2014), pp. 122-131 © IAEME 131 [8] Disha Sharma, Gagandeep Jindal. Identifying Lung Cancer Using Image Processing Techniques. International Conference on Computational Techniques and Artificial Intelligence (ICCTAI'2011). [9] Mrs.Mamata S.Kalas, Artificial Neural Network for Detection of Biological Early Brain Cancer. ©2010 International Journal of Computer Applications (0975 – 8887) Volume 1 –No. 6. [10] Saif D. Salman & Ahmed A. Bahrani Segmentation of tumor tissue in gray medical images using watershed transformation method. International Journal of Advancements in Computing Technology Volume 2, Number 4, October 2010. [11] Manoj K Kowar and Sourabh Yadav, Brain Tumor Detction and Segmentation Using Histogram Thresholding. International Journal of Engineering and Advanced Technology (IJEAT) ISSN: 2249 – 8958, Volume-1, Issue-4, April 2012. [12] Mukesh Kumar, Kamal K.Mehta, A Texture based Tumor detection and automatic Segmentation using Seeded Region Growing Method. Mukesh Kumar et al, Int. J. Comp. Tech. Appl., Vol 2 (4), 855-859 IJCTA | JULY-AUGUST ISSN:2229-6093. [13] P.K.Srimani and Shanthi Mahesh. A Comparative Study of Different Segmentation Techniques for Brain Tumor Detection. International Journal of Emerging Technologies in Computational and Applied Sciences (IJETCAS). [14] B.Venkateswara Reddy, Dr.P.Satish Kumar, Dr.P.Bhaskar Reddy and B.Naresh Kumar Reddy, “Identifying Brain Tumour from MRI Image using Modified FCM and Support Vector Machine”, International Journal of Computer Engineering & Technology (IJCET), Volume 4, Issue 1, 2013, pp. 244 - 262, ISSN Print: 0976 – 6367, ISSN Online: 0976 – 6375. [15] Selvaraj.D and Dhanasekaran.R, “MRI Brain Tumour Detection by Histogram and Segmentation by Modified GVF Model”, International Journal of Electronics and Communication Engineering & Technology (IJECET), Volume 4, Issue 1, 2013, pp. 55 - 68, ISSN Print: 0976- 6464, ISSN Online: 0976 –6472. [16] Mayur V. Tiwari and D. S. Chaudhari, “An Overview of Automatic Brain Tumor Detection From Magnetic Resonance Images”, International Journal of Advanced Research in Engineering & Technology (IJARET), Volume 4, Issue 2, 2013, pp. 61 - 68, ISSN Print: 0976-6480, ISSN Online: 0976-6499.