A Dualistic Sub-Image Histogram Equalization Based Enhancement and Segmentation Techniques with NN for Medical Images
•
1 like•729 views
This document presents a dualistic sub-image histogram equalization technique for medical image enhancement and segmentation. The technique divides an image histogram into two parts based on mean and median, then equalizes each sub-histogram independently. It enhances images effectively while constraining average luminance shift. For segmentation, canny edge detection and neural networks are used. The technique is tested on medical images and shows improved completeness and correctness over previous methods, with neural networks increasing accuracy to 98.3257%.
Report
Share
A Dualistic Sub-Image Histogram Equalization Based Enhancement and Segmentation Techniques with NN for Medical Images
- 1. Research Inventy: International Journal Of Engineering And Science Vol.05, Issue 01 (January 2015), PP: 15-19 Issn (e): 2278-4721, Issn (p):2319-6483, www.researchinventy.com 15 A Dualistic Sub-Image Histogram Equalization Based Enhancement and Segmentation Techniques with NN for Medical Images 1 Mandeep Kaur, 2 Ishdeep Singla 1 Research Scholar Department of Computer Science , Chandigarh University Gharuan, India 2 Assistant Professor Department of Computer Science Chandigarh University Gharuan, India ABSTRACT:-Histogram Equalization is a contrast enhancement technique in the image processing which uses the histogram of image. Segmentation of image plays a vital role in many medical imaging applications by automatically locating the regions of interest. Segmentation of image is the more crucial functions in figure analysis and processing. The segmentation results of a figure affect all the subsequent processes of image analysis. This is necessary to develop medical figure segmentation algorithms that are accurate and efficient. In this work; develop a dualistic sub-image histogram equalization based enhancement and segmentation techniques. Dualistic sub image histogram equalization (DSIHE) which divides the image histogram into two parts based on the input mean and median respectively then equalizes each sub histogram independently. Further to enhance work, use NN to present better result as compare to previous work. The proposed method has been tested and evaluated on several medical images. In paper, the medical figure is lineated and extracted out so that it can be viewed individually. Then results demonstrate that the developed algorithm is highly efficient over hierarchical grouping technique. It is valid using the performance measures such as completeness and clearness. For the implementation of this proposed work; use the GUI and NN Toolbox under Matlab software Keywords: - Histogram Equalizations; Contrast Enhancement; Brightness Preservation; Absolute Mean Brightness Error; Peak Signal to Noise Ratio; Structure Similarity Index and Image Processing I. INTRODUCTION In medical imaging there is a massive amount of information, but it is not possible to access or make use of this information if it is efficiently organized to extract the semantics. To retrieve semantic image, is a hard problem. In image retrieval and pattern recognition community, each image is mapped into a set of numerical or symbolic attributes called features, and then to find a mapping from feature space to image classes. Image classification and image retrieval share fundamentally the same goal if there is given a semantically well- defined image set. Dividing the images which is based on their semantic classes and finding semantically similar images also share the same similarity measurement and performance evaluation standards. An image retrieval framework consisting of three stages; feature extraction, feature selection and image retrieval using k-nearest neighbors in the selected feature space. Neurology is the current focus of the knowledge bank. These images are scanned from the CT or MRI. Medical image segmentation is the method of labeling each voxel in a medical image dataset to state its anatomical structure. The labels that result from this method have a wide variety of applications in medical research. Segmentation is a very common method so it is difficult to list most of the segmented areas, but a general list would consists of at least the following; the brain, heart, knee, jaw, spine, pelvis, liver, prostate, and the blood vessels. The input to a segmentation process is grayscale digital medical image, (like CT or MRI scan). The desired output restrains the labels that classify the input grayscale voxels. The use of segmentation is to give preeminent information than that which exists in the original medical images only. The set of labels that is produced through segmentation is also called a label map, which briefly tells its function as a voxel by voxel guide to the original imagery. Therefore frequently used to improve visualization of medical figure and allow quantitative measurements of image structures; segmentation are also important in building anatomical atlases; researching shapes of anatomical structures; and tracking anatomical changes over time. Segmentation of medical images is a challenging issue in the field of image processing. Several literatures exist in segmentation of medical images. In our proposed technique we use the DSIHE for the image enhancement and after that apply the segmentation. After all previous techniques used; further use NN. The flowchart of the proposed segmentation algorithm is shown in Figure1. The details of each step are explained in the following section.
- 2. A Dualistic Sub-Image Histogram... 16 Figure1: Step of proposed work II. INPUT IMAGE In our proposed approach we first considered that the MRI scan images of a given patient are either color; the Gray-scale or intensity figures herein are displayed with a default size. If it is color image, a Gray- scale converted image is defined by using a large matrix whose entries are numerical values between zero and 255; where zero corresponds to black and 255 to white for instance. Thus the brain tumor detection of a given patient consists of two main stages namely; image segmentation and edge detection III. IMAGE SEGMENTATION The objective of image segmentation is to cluster pixels into prominent image region. This paper; segmentation of Gray level figures is used to provide information such as anatomical structure and identifying the Region of Interest i.e. locate tumour, lesion and other abnormalities. The propose technique is depend upon the information of anatomical structure of the healthy parts and compares it with the infected parts. This begins by allocating the anatomical structure of the healthy parts in a reference image of a normal candidate brain scan image. In this paper for segmentation; use canny edge detection which is discussed below: Canny Edge Detection: The Canny edge detector is an edge detection operator to detect a wide range of edges in images. The algorithm runs in 5 separate steps: - I. smoothing: - Blurring of the image to remove noise. II. Finding gradients: - The edges should be marked where the gradients of the image has large magnitudes. III. Non-maximum suppression: - Only local maxima should be marked as edges. IV. Double Thresholding: - Potential edges are determined by Thresholding. V. Edge tracking by hysteresis: - Final edges are determined by suppressing all edges that are not connected to a very certain (strong) edge. IV. HISTOGRAM EQUALIZATION Histogram equalization is a method in image processing of contrast adjustment using the image's histogram. It is a method usually increases the global contrast of many images; especially when the usable data of the image is represented by close contrast values. To through this adjustment; the intensities can be better distributed on the histogram. It is allowed for area of lower local contrast to gain a more contrast. The histogram equalization is accomplishes; this by effectively spreading out the most frequent intensity values. V. DUALISTIC SUB-IMAGE HISTOGRAM EQUALIZATION This is a novel histogram equalization technique in which the original image is decomposed into two equal area sub-images based on its gray level probability density function. Therefore two sub-images are equalized respectively. At end; get the result after the processed sub-images are composed into one image. The fact that algorithm can not only enhance the image visual information effectively; but also constrain the original image's average luminance from great shift. It makes this possible to be utilized in video system directly.
- 3. A Dualistic Sub-Image Histogram... 17 VI. HOLE FILLING A hole is a background region surrounded by a connected border of foreground elements. An image that could result from thresholding to 2 levels a scene containing polished spheres (ball bearings). Dark spots could be results of reflections. Objective is to remove reflections by hole filling. The holes in the segmented image are due to noise. To eradicate this, we invert the segmented image. That is the pixels are labeled as 0‟s and non-image pixels are labeled as 1‟s. VII. BRANCH REMOVAL By using „bwmorph‟, one can identify the branch points in a skeleton and by subtracting them from the skeleton image; the residual image is left with the different branches. Most image editors can be used to remove unwanted branches, etc., using a "clone" tool. Removing these distracting elements draws focus to the subject; improving overall composition. VIII. NEURAL NETWORK Artificial neural networks are composed of interconnecting artificial neurons (programming constructs that mimic the properties of biological neurons). Hence, Artificial neural networks may either be used to gain an understanding of biological neural networks; or for solving artificial intelligence problems without necessarily creating a model of a real biological system. Hence, real; biological nervous system is highly complex: artificial neural network algorithms attempt to abstract this complexity and focus on what may hypothetically matter most from an information processing point of view. Good performance (e.g. as measured by good predictive ability; low generalization error); or performance mimicking animal or human error patterns; can then be used as one source of evidence towards supporting the hypothesis that the abstraction really captured something important from the point of view of information processing in the brain. Other incentive for these abstractions is to reduce the amount of computation required to simulate artificial neural networks; so as to allow one to experiment with larger networks and train them on larger data sets. IX. RESULT DISCUSSION Figure 2: Graph of correctness between Brain and algorithm
- 4. A Dualistic Sub-Image Histogram... 18 Figure 3: Graph between completeness between brain and algorithm Table 1: Performance Evaluation of Brain images Images Completeness Correctness Brain Previous propose previous Propose 0.944 5.9 0.956 3.6 The table show the value of correctness and completeness of previous technique and propose technique. And the value completeness and correctness of propose techniques much better than previous technique as shown in table. Figure 4: value of Accuracy by using NN technique The above figure shows the result of work done. Here use DSIHE for segmentation and equalization of medical image. To enhance the previous work; use NN technique to give better result as compare to previous result. Here accuracy reaches to 98.3257% by using NN. X. CONCLUSIONS This paper will present a complete and fully segmentation and enhancement of medical image by using DSIHE technique. Here we apply enhance the image after that segmentation done on that image. After these processes; hole filling and branch removal is done respectively on the medical image. To further enhance work and produce better result use NN technique. The NN algorithm enhance both parameter value i.e. correctness and completeness and also show better accuracy as compare to previous technique.
- 5. A Dualistic Sub-Image Histogram... 19 REFERENCES [1]. Naeem Shareef, DeLiang L. Wang “Segmentation of Medical Images Using LEGION”, IEEE Transactions On Medical Imaging, Vol. 18, No. 1, January 1999. [2]. Julio Carballido-Gamio, Serge J. Belongie, and Sharmila Majumdar “Normalized Cuts in 3-D for Spinal MRI Segmentation”, IEEE Transactions on Medical Imaging, Vol. 23, No. 1, January 2004. [3]. Smadar Shiffman*, Geoffrey D. Rubin, and Sandy Napel “Medical Image Segmentation Using Analysis of Isolable-Contour Maps”, IEEE Transactions On Medical Imaging, Vol. 19, No. 11, November 2000. [4]. Naeem Shareef, DeLiang L. Wang “Segmentation of Medical Images Using LEGION”, IEEE Transactions On Medical Imaging, Vol. 18, No. 1, January 1999. [5]. Myungeun Lee, Chonnam, Gwangju, Wanhyun Cho,Soohyung Kim, “Segmentation of medical images using a geometric deformable model and its visualization” ”, IEEE Transactions On Medical Imaging, Vol. 15, No. 2, February 2001. [6]. Chen X, Udupa JK, Bagci U, Zhuge Y, Yao J., “Medical Image Segmentation by Combining Graph Cuts and Oriented Active Appearance Models”, IEEE Trans Image Process. 2012. [7]. Jianfeng Xu, Lixu Gu, Xiahai Zhuang, and Terry Peters., “A Novel Multistage 3D Medical Image Segmentation: Methodology and Validation”, Springer-Verlag Berlin Heidelberg 2005, pp. 884 – 889. [8]. Pingkun Yan and Kassim, A.A.,”Medical Image Segmentation Using Minimal Path Deformable Models With Implicit Shape Priors”, IEEE Transactions on Information Technology in Biomedicine, Vol.10, Iss.4, Oct. 2006, pp. 677 – 684. [9]. Bomans M, Hohne KH, Tiede U, Riemer M. “3-D segmentation of MR images of the head for 3-D display”, IEEE Trans Med Imaging. 1990, Vol. 9, No. 2, pp.177-183. [10]. D. Chaudhuri, N. K. Kushwaha, and A. Samal, “Semi-Automated Road Detection from High Resolution Satellite Images by Directional Morphological Enhancement and Segmentation Techniques”, IEEE Journal of Selected Topics In Applied Earth Observations And Remote Sensing, Vol. 5, No. 5, October 2012. [11]. D. Chaudhuri, C. A. Murthy, and B. B. Chaudhuri, “A modified metric to compute distance,” Pattern Recognition, Vol. 25, No. 7, pp. 667–677, 1992. [12] Pankaj Sapra, Rupinderpal Singh, Shivani Khurana”Brain Tumor Detection Using Neural Network”International Journal of Science and Modern Engineering (IJISME) ISSN: 2319-6386, Volume-1, Issue-9, August 2013 [13] 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 16 [14] M.G. Masoole and A. S. Moosavi. “An Improved Fuzzy Algorithm for Image Segmentation”. World Academy of Science, Engineering and Technology Vol. 38. 2008. Pp. 400-404. [15] L. Jiang and W. Yang. “A Modified Fuzzy C-Means Algorithm for Segmentation of Magnetic Resonance Images”. Proc. VIIth Digital Image Computing: Techniques and Applications. Sydney 10-12 Dec. 2003. [16] D.L. Pham and J.l. Prince. “Adaptive fuzzy segmentation of magnetic resonance images”. IEEE Trans. in Medical Imaging. 1999. Vol. 18. pp. 737–752. [17] R.Gonzalez and R. Woods, Digital image processing using MATLAB 5th ed. Pearson , 2009 . [18] Y.-T. Kim, Contrast enhancement using brightness preserving bi histogram equalization, “IEEE Trans. On Consumer Electronics, vol. 43, no. 1, pp. 1-8, Feb. 1997. [19] David Menotti, Laurent Najman, Jacques Facon, “ Multi-histogram equalization methods for contrast enhancement and brightness preserving”, “IEEE Trans. On Consumer Electronics, vol. 53, no. 3, Aug. 2007”. [20] P. Jagatheeswari, S. Suresh Kumar and M. Rajaram, “ A novel approach for contrast enhancement based on histogram equalization followed by median filter”, “ARPN Journal of Engineering and Applied Sciences, Vol. 4 , No. 7, Sep. 2009. [21] M. Karaman, L. Onural, and A. Atalar, “ Design and implementation of a general purpose median filter unit in CMOS VLSI “, “ IEEE Trans.” April. 1990.