Tumor Detection In Prostate Organ Using Canny Edge .

3y ago
56 Views
2 Downloads
310.15 KB
8 Pages
Last View : 1d ago
Last Download : 3m ago
Upload by : Brenna Zink
Transcription

International Journal of Pure and Applied MathematicsVolume 118 No. 9 2018, 211-217ISSN: 1311-8080 (printed version); ISSN: 1314-3395 (on-line version)url: http://www.ijpam.euSpecial Issueijpam.euTumor Detection in Prostate Organ UsingCanny Edge Detection TechniqueN. GopinathAssistant ProfessorDepartment of Computer Science and EngineeringSri Sai Ram Engineering College, Chennai, Indiagopinath.cse@sairam.edu.in, bajjugopi@gmail.comAbstract— Medical imaging is a technique and process of creating visual representations of the interior of a bodyfor clinical analysis and medical intervention, as well as visual representation of the function of some organs ortissues. It seeks to reveal internal structures hidden by the skin and bones, as well as to diagnose and treat disease.Magnetic Resonance Imaging (MRI) is a medical imaging technology. This MRI uses radio waves and also amagnetic field to create detailed images of organs and tissues. This MRI has proven to be highly effective indiagnosing a number of conditions by showing the difference between normal and diseased soft tissue of the body.In recent years, this MRI imaging has emerged as an alternative to other Imaging modalities like US, PET and so onfor clear identification of cancer cells which cause cancer disease in Human Breast, Prostate and liver etc. Toanalyze a disease more effectively physicians consider MR Imaging modality for identification of cancer cells invarious organs. Therefore, analysis on MR imaging is required for efficient diagnosis of diseases. One of themethod used for detecting tumor / cancer cells in Prostate is based on edge based segmentation for which we areusing Canny Edge Detection Technique. This proposed canny Edge detection technique is a part of Image Analysis.This proposed work is based on the following procedures: 1. Image Enhancement 2. Remove Artifacts 3. ApplyCanny Edge Detection technique. This method was implemented using Matlab Code.Keywords—Magnetic Resonance Imaging (MRI); Prostate Cancer; Canny Edge Detection; Region Of Interest(ROI)1. IntroductionCancer is defined as a group of diseases involving abnormal cell growth with the potential to invade or spread toother parts of the body. In short cancer is called as an abnormal growth of tissue cells. Prostate Cancer also knownas Carcinoma of the Prostate, is the development of cancer in the Prostate, a gland in the male reproductive system.Most Prostate cancers are slow growing [1]. The cancer cells may spread from the Prostate organ to the other partsof the body; particularly, the bones and lymph nodes. Rates of detection of Prostate cancer vary widely across theworld [2]. This Prostate cancer is the second leading cause of cancer – related death in the United States among menand it is the most commonly diagnosed cancer in American males [3]. The early detection or identification ofProstate cancer plays a vital role in the success of treatment and outcome. To detect Prostate cancer, imagingmodalities like Trans Rectal Ultra-Sound (TRUS) and Magnetic Resonance Imaging (MRI) are relied on. In thisproposed technique we have chosen MRI modality of image for detection of tumor in Prostate organ. The reason forchoosing MRI modality is that, it always produce a high quality images of the parts contained in the human body.From these high resolution images, we can derive detailed, anatomical information to examine humanorgandevelopment and also to discoverabnormalities. Image Processing is termed as a technique to enhance rawimages received from cameras / sensors placed on satellites, or pictures taken in normal day –to- day life for variousapplications. It is significantly improved in recent times and extended to various fields of science and technology.The image processing mainly deals with the Image acquisition, Image enhancement, Image segmentation, featureextraction, image classification etc. [4]. Edge detection technique is one of an Image processing technique for211

International Journal of Pure and Applied MathematicsSpecial Issuefinding the boundaries of objects within images. This technique works by detecting discontinuities in brightness. Itis used for Segmentation of Images and extraction of data in areas such as Image processing, computer vision andmachine vision. Several Edge detection algorithms like Sobel, Canny, Prewitt and fuzzy logic are exists, and in ourproposed method we used on a particular one named “Canny Edge method”, developed by John F. Canny (JFC) [5].Even though it is quite old, it has become one of the standard edge detection methods and it is still used in research.Canny Edge detection algorithm runs in 5 steps. They are 1. Smoothing, 2. Finding Gradients, 3. NonMaximumSuppression, 4. Double Thresholding and 5. Edge tracking by hysteresis. The algorithm has been tried on number ofpatients MRI data of Prostate cancer images [5]. The organization of the paper is as follows: Section II involves thetechniques being used in this paper to extract the tumor in the Prostate. Section III contains the experimental resultsof the techniques that are described in this paper and section IV consists of conclusion and also the informationrelated to the future work of this proposed method.2. Proposed MethodologyThe System design of the proposed methodology is shown in figure 1:Image EnhancementRemove ArtifactsApply Canny Edge Detection TechniqueFig 1. Flowchart showing the steps of the proposed methodologyA. Data setApproximately 70% of the prostate is composed of glandular tissue, and 30% consist of non-glandular tissue.For anatomic division of the prostate, the zonal compartment system developed by Mc. Neal is widely accepted [6],[7]. According to this system, glandular tissue is sub divided into the central and the peripheral gland. The central glandis composed of a transitional zone and Peri urethral tissue, and the peripheral gland is composed of peripheral andcentral zones (Fig 2). The peripheral zone includes the posterior and lateral aspects of the Prostate. This peripheral zoneaccounts for most of the glandular tissue. 70% of Prostate cancer arise in this zone. The transitional zone accounts for5% of the glandular tissue of the Prostate. Benign prostatic hyperplasia arise as a result of Cellular proliferation in thetransitional zone. In addition, 20% of the Prostate cancer arise in the transitional zone.Fig 2. Schematics shows the anatomy of the Prostate.The abbreviations for the terms mentioned in the above figure 2 are given as, AFT Anterior Fibro muscularTissue, TZ Transitional Zone, PUT Peri Urethral Tissue, CZ Central Zone, PZ Peripheral Zone.B. Grayscale ImagingMR Images or Magnetic Resonance Images, which can be acquired on computer when a patient is scanned byMRI machine. We can also acquire MR Images of the part of the body which is under test or desired. Generally MRImages on computer are looks like black and white images. In analog practice, the gray scale imaging is sometimescalled as “Black and White”, but technically this is a misnomer. In true, this black and white images are also known ashalftone, the only possible shades are pure black and pure white. A grayscale (or gray level) image is one in which theonly colors are shades of gray. The main reason for differentiating such images from any other sort of color image is212

International Journal of Pure and Applied MathematicsSpecial Issuethat less information needs to be provided for each pixel. In general a gray color is one in which the red, green and bluecomponents all have equal intensity in RGB space and so it is only necessary to specify each pixel in a full color image.Grayscale is said to be a range of shades of gray without apparent color. The darkest possible shade is black, whichmeans the total absence of transmitted or reflected light. Similarly the lightest possible shade is white, which meansthat the total transmission or reflection of light at all visible wavelengths. So because of the above reason first weconvert our MR Image to be pre-processed in grayscale image.C. Image EnhancementIn Image enhancement, the main aim is to accentuate certain image features for subsequent analysis or forimage display. Few examples are contrast and edge enhancement, pseudo coloring, noise filtering and magnifying.Image enhancement is useful in feature extraction, image analysis and visual information display [8]. This ImageEnhancement plays a significant role in Image processing at low level. The main purpose of Preprocessing is to enlargethe intensity or brightness difference between objects and background and to produce more proper and reliablerepresentations of Prostate tissue structures. Here enhancement is performed because of low contrast of MRI Prostateimages and variations of intensity of objects with background.D. Remove ArtifactsRemove artifacts includes the removal of background information such as labels and text in it. We perform thefollowing steps:1. Set a threshold value in order to convert gray scale image to binary image. Here the threshold value we fixedis 20 which is obtained by trial and hit method.2. After getting the binary image, our next move is based on morphological operations which are erosion anddilation. In order to do the morphological operations on the image, the structuring element we are taking is diskhaving radius of 8 pixel with 8 approximations. At this stage all binary objects are cleaned except the Prostateregion.E. Canny Edge Detection TechniqueAn Edge in an image is a significant local change in the image intensity, normally associated with adiscontinuity in image intensity or in the first derivative of the image intensity. [9]. Edge detection refers to the processof identifying and location sharp discontinuities in an image. The discontinuities are termed as the abrupt changes inpixel intensity which characterize boundaries of objects in a scene. Most classical methods of edge detection involvesconvolving the image with an operator (a 2D filter) which is erected to be sensitive to large gradients in the imagewhile returning values of zero in uniform regions. There are an extremely large numbers of edge detection operatorsavailable, each designed to be sensitive to certain types of edges. Variables which are involved in the selection of anedge detection operator include edge orientation, Noise environment and Edge structure [10].There are many edge detection techniques in the literature of Image segmentation. The most commonly useddiscontinuity based edge detection techniques are Roberts’s edge detection, Sobel edge detection, Prewitt edgedetection, Kirsh edge detection, Robinson edge detection, MarrHildreth edge detection, LoG edge detection and CannyEdge detection [11]. Here, in this proposed method we are using canny operator for edge detection as it providesstronger edges as compared to other edge detectors.The canny edge detection technique includes the following steps which are given in flowchart shown below:Gaussian SmoothingFiltering of GradientNon Maximal SuppressionHysteresis ThresholdingFig 3. Flowchart showing different steps of Canny Edge Detection technique213

International Journal of Pure and Applied MathematicsSpecial IssueI. Gaussian SmoothingGaussian Smoothing is the first step of canny edge detection method. Here, de-noising of an image isperformed with the help of a 2D Gaussian filter. Noise in an image may lead to un-intended edges, so de-noising isperformed. The Gaussian smooth filter is one of the best filter for removing noise drawn from normal distribution.Gaussian filters are a class of linear smoothing filters with the weight chosen according to the shape of a Gaussianfunction. For image processing, the two dimensional zero –mean discrete function, g[i,j] e-(i2 j2) / 2σ2 is used as aSmoothing filter [12]. The quality of smoothing is depends on the value of sigma (Standard deviation) what we haveused. The width of the Gaussian filter is defined through this sigma. This sigma value should neither be too small norbe too large and most importantly this value should be a positive value. The negative sigma value indicates theobservation through infinitesimally small area which is impossible. 0.5 is the smallest value of Sigma.II. Gradient FilteringBy using the Gaussian Smoothing method, the noise present in the given input image are eliminated, the next stepis to find the edge strength by taking the gradient of the image. In image processing, a gradient refers to the distributionof the brightness throughout the image. Gradient values are not only the measure forbrightness but also for thedirectionof edges. Edges are normal to the gradients. The series of pixel at which there are abrupt changes in gradient valuefrom pixel to pixel indicates the presence of edge.Edge strength can be find by using Sobel operator. This Sobel operator performs a 2-D spatial gradientmeasurement on the de-noised image. Then, the approximate absolute gradient magnitude (edge strength) at each pointcan be found after measuring the spatial gradient value. The Sobel operator uses a pair of 3*3 convolution masks, oneestimating the gradient in the X-direction (columns) and other estimating the gradient in the Y-direction (rows) [13].The magnitude or edge strength of the gradient is then approximated using the below mentioned formula: GR GRx GRy (1)Finding the edge direction is unimportant. Once the gradient in the X and Y directions are known. The formula forfinding the edge direction is given below:Θ invtan (GRy / GRx)(2)Once the edge directions are known, non-maximum suppression now has to be applied.III. Non maximal SuppressionThe edges obtained in the gradient image are thick, due to which it becomes very difficult to localize the edgesproperly. So, to have thin edges we have to perform the next step called as Non maximal suppression. In Non-maximalsuppression, the pixels having the maximum value of gradient are preserved and the pixels at which the gradient valuesis not maximum are suppressed. It means they are set to zero.IV. Hysteresis ThresholdingThe suppressed pixel values leads to the partial filtering of relevant pixels, so in order to suppress theremaining irrelevant pixels, we perform the next step called as Thresholding. The thresholding is defined as a processof converting the grayscale image into a bi-level image using an optimal threshold. The purpose of thresholding is toextract those pixels from some images which represents an objects [14].The thresholder used in the canny operator iscalled as “Hysteresis”. Most thresholder uses a single threshold limit, which means if the edge value fluctuate aboveand below this value the line will appear broken (which is commonly called as “Streaking”). Hysteresis countersstreaking by setting an upper and lower edge value limit. Considering a line segment, if a value lies above the upperthreshold limit, then it is immediately accepted. If the value lies below the low threshold it is then immediatelyrejected. Points which lie between the upper and lower limits are accepted, if they are connected to pixels which exhibitstrong response. The likelihood of streaking is reduced drastically since the line segment points must vary above theupper limit and below the lower limit for streaking to occur [15].In this proposed work, we are using the below mentioned expression to know which threshold value will becalculated dynamically at run time for each image:High Threshold min (fin (cumsum (counts) PPNE*m*n)) /64(3)Where, cumsum is Cumulative sum and PPNE is the percentage of pixels which are not taking part in the edgeformation. Here, the value we are taking is 0.7. The value for low threshold is found by multiplyingthe threshold ratiowith high threshold value. The threshold ratio we are using is 0.4 for each Prostate MRI image. Hysteresis thresholdingis performed in order to have unstreak edges and helps to trace the true edges.V. Filling of HolesAfter having the edges through Hysteresis thresholding, we are going to perform one kind of Morphologicaloperation called as “Filling of holes”. A hole is defined as a set of background pixels that cannotbe reached by filling inthe background from the edge of the image. In order to fill the set of background pixels, we look for theclose loops214

International Journal of Pure and Applied MathematicsSpecial Issueandthose close loops are highlighted by using an in built Matlab command named “Imfill‟ that fills all the close loopspresent in the binary image BW.VI. Perform Dilation and Erosion OperationThe most basic morphological operations are the Dilation and Erosion. Dilation is for adding pixels to theboundaries of objects in an image, while erosion is for removing pixels on object boundaries. The number of pixelsadded or removed from the object in an image depends on the size and shape of the structuring element used to processthe image. In this morphological dilation and erosion operations, the state of any given pixel in the output image isobtained by applying a rule to the corresponding pixel and its neighbors in the input image. The rule used to process thepixel defines the operation as a dilation or an erosion. The rule to perform dilation operation is given as: The value ofthe output pixel is the maximum value of all the pixels in the input pixel’s neighborhood. In a binary image, if any ofthe pixel is set to the value1, the output pixel is set to 1. Similarly, for erosion the rule is given as: The value of theoutput pixel is the minimum or lowest value of all the pixels in the input pixel’s neighborhood. In a binary image, ifany of the pixels is set to 0, the output pixel is set to 0.Erosion removes small-scale details from a binary image but simultaneously reduce the size of Regions ofInterest (ROI), too. So in order to restore the original shape of ROI, we perform dilation.3. EXPERIMENTAL RESULTSTo implement our proposed method, real time patient data is taken for analysis. As cancer in MRI image have highintensity when compare to its background so it becomes very easy to identify it from a MRI image. The result obtainedfrom each module of our proposed method is shown here:Fig 4. Results of each module from the proposed methodIn the above mentioned figure 4, the captions from a to j is defined as follows: (a) Input MRI Prostate image withcancer affected region- Highlighted (b) Background subtracted image from the original image (c) Contrast increasedimage (d) Gray scale to Binary converted image (e) Binary image with noise (f) De-noised image using Gaussian 2DFilter (g) Sobel gradient image (h) Thin edged image (i) Close loop filled image (j) Tumor detected image.Conclusion and Future worksFrom the result of this proposed method we can say that the highlighted cancer affected region in our inputimage is detected exactly with the help of Canny Edge Detection technique. This Canny Edge Detection technique is215

International Journal of Pure and Applied MathematicsSpecial Issueused as a pre-processing step in order to extract the tumor. It provides the strong edges as compared to other edgedetectors and has overcome the problem of threshold adjustment, by which we can extract the tumor or cancer by usingmorphological operations more easily and appropriately.For future work, this proposed algorithm can be extended for some other modality of images like CT, UltraSound, etc., for different cancer affected organs of human such as breast, brain and so on. This proposed system findsits application in the Medical field and other research areas.References[1]Rajesh C. Patil and Dr.A.S. Bhalchandra, “Brain Tumor Extraction from MRI Image Using MATLAB”, IJECSCSE, ISSN:2277-9477, Volume 2, Issue 1.[2]Sam Lister (February 11, 2009). “Rine test could speed treatment of Prostate Cancer”, London. The Sunday Times. Retrived 9August 2010.[3] Olga Milikovi, “ Image pre-processing Tool”, College of Computer Science, Megatrend University of B

performed with the help of a 2D Gaussian filter. Noise in an image may lead to un -intended edges, so de -noising is performed. The Gaussian smooth filter is one of the best filter for removing noise drawn from normal distribution. Gaussian filters are a class of linear smoothing filters with the weight chosen according to the shape of a Gaussian

Related Documents:

Plus, prostate problems are not an old man’s disease any more. More and more men are dealing with prostate disease at ever younger ages. Western males are particularly at risk and are 30 to 50 times more likely of getting prostate cancer than an Asian, Indian or African man. Worse if you are African American, who have the highest rates in the .

8 Understanding the PSA test: A guide for anyone concerned about prostate cancer Urinary problems and prostate problems If you notice any changes when you urinate or have any urinary problems (see below), it could be a sign of a prostate problem. Urinary problems are common in older men and aren't always a sign of a prostate problem.

Advanced prostate cancer is when the cancer is no longer contained within the prostate gland, and cancer cells have spread to other parts of the body. There are different stages of advanced prostate cancer: Locally advanced - the cancer has extended beyond the prostate and may include seminal vesicles (tumour stage T3) or other

single tumor. Tumor; mass; tumor mass; lesion; neoplasm o The terms tumor, mass, tumor mass, lesion, and neoplasm are . not. used in a standard manner in clinical diagnoses, scans, or consults. Disregard. the terms unless there is a . physician's statement . that the term is malignant/cancer o These terms are used . ONLY to determine .

Hepatocellular carcinoma (HCC) is one of the most common malignant tumors worldwide and is associated with poor clinical prognosis, which is mainly caused by tumor recurrence and metastasis. Circulating tumor cells (CTCs) are tumor cells shed into the bloodstream and regarded as the “seeds” of tumor recurrent or metastatic lesions.

Identify the organ systems in Figure 2.1 using the organ system list provided. Refer to Table 2.1. A. Overview of Organ Systems The body stays alive due to the interaction of different organ systems. An organ systemis a group of organs p

990 CD Sheet Music 984 Church Organ Folios 982 Organ Adventure 982 Organ Instruction . les MisÉraBles 14 songs from Broadway’s longest-running musical arranged for organ, . I Dreamed a Dream In My Life On My Own and more. _00290270. 12.99 105 favorite hyMns organ

image processing. Detection and extraction of tumor from MRI scan images of the brain is done by using MATLAB software. The aim of this work is to design an automated tool for brain tumor quantification using MRI image datasets. Key words: Brain tumor, grey scale imaging, MRI, MATLAB, morphology, noise removal, segmentation .