11.[37 46]segmentation and feature extraction of tumors from digital mammograms
1 like541 views
This document summarizes a proposed computer-aided detection (CAD) system for segmenting tumors and extracting features from digital mammograms to help radiologists diagnose breast cancer. The proposed CAD system includes five stages: 1) collecting input images and extracting regions of interest, 2) enhancing the regions of interest, 3) segmenting the tumors, 4) filtering noises, and 5) extracting texture and statistical features. Texture features like Haralick features extracted from gray-level co-occurrence matrices are calculated to classify tumors as benign or malignant. The system is tested on mammogram images from the Mammographic Image Analysis Society database to evaluate the segmentation and feature extraction methods.
![Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.4, 2012
Segmentation and Feature Extraction of Tumors from Digital
Mammograms
PRADEEP N *,
Department of Computer Science and Engineering, Bapuji Institute of Engineering and Technology,
Davangere, Karnataka, INDIA.
Email: nmnpradeep@yahoo.com
GIRISHA H,
Department of Computer Science and Engineering, Rao Bahadur Y Mahabaleswarappa Engineering
College, Bellary, Karnataka, INDIA.
KARIBASAPPA K,
Department of Computer Science and Engineering, Dayananda Sagar College of Engineering,
Bangalore, Karnataka, INDIA.
Abstract
Mammography is one of the available techniques for the early detection of masses or abnormalities which
is related to breast cancer. Breast Cancer is the uncontrolled of cells in the breast region, which may affect
the other parts of the body. The most common abnormalities that might indicate breast cancer are masses
and calcifications. Masses appear in a mammogram as fine, granular clusters and also masses will not have
sharp boundaries, so often difficult to identify in a raw mammogram. Digital Mammography is one of the
best available technologies currently being used for the early detection of breast cancer. Computer Aided
Detection System has to be developed for the detection of masses and calcifications in Digital
Mammogram, which acts as a secondary tool for the radiologists for diagnosing the breast cancer. In this
paper, we have proposed a secondary tool for the radiologists that help them in the segmentation and
feature extraction process.
Keywords: Mammography, Breast Cancer, Masses, Calcification, Digital Mammography, Computer Aided
Detection System, Segmentation, Feature Extraction
1. INTRODUCTION
Cancer is an abnormal, continual multiplying of cells. The cells divide uncontrollably and may grow into
adjacent tissue or spread to distant parts of the body. Breast cancer remains a leading cause of cancer deaths
among women in many parts of the world. Early detection of breast cancer through periodic screening has
noticeably improved the outcome of the disease [1]. The mass of cancer cells will eventually become large
enough to produce lumps, masses, or tumors that can be detected. Tumor is uncontrolled growth of cells
which can be either Benign or Malignant. Benign Tumors are not cancerous. Benign tumors may grow
larger but do not spread to other parts of the body. Malignant Tumors is cancerous. Malignant tumors can
invade and destroy nearby tissue and spread to other parts of the body. Tumor can be easily identified in
mammogram because tumor part is highly bright (having high intensity) compared to other part
37](https://image.slidesharecdn.com/11-37-46segmentationandfeatureextractionoftumorsfromdigitalmammograms-120512235750-phpapp02/85/11-37-46-segmentation-and-feature-extraction-of-tumors-from-digital-mammograms-1-320.jpg)
![Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.4, 2012
(background) of the mammogram image as shown in Figure 1.
In the figure 1, we can observe that the marked oval shape area have higher intensity compared to the
surrounding area. This marked oval shape is the required Region Of Interest (ROI). Segmentation
techniques can be used to extract the tumor from the mammogram. Computer Aided (CA) detection
systems have been developed to aid radiologists in detecting mammographic abnormalities [2-5].
There are large numbers of diagnostic methods currently available for the diagnosis of Breast Cancer,
among which digital mammography is the most reliable method, for detecting early breast cancer [6-7].
Early detection of Breast Cancer can decrease the mortality rate. In this paper, we have proposed a CAD
system to segment the tumor from the ROI and calculate the features that have importance in the
classification of tumors.
2. PROPOSED CAD SYSTEM
The proposed CAD system consists of five principal stages. The first stage is the collection of input images
and ROI extraction, second stage is ROI Image Enhancement, the third stage is Segmentation, the fourth
stage is Filter and the last stage is Feature Enhancement.
2.1 ROI Extraction
The input images are collected from Mammographic Image Analysis Society (MIAS) research
database. The MIAS dataset [6] is used to test the proposed technique. These images were previously
investigated and labeled by an expert radiologist based on technical experience and biopsy. The dataset is
selected due to the various cases it includes. The dataset is composed of 322 mammograms of right and left
breast, from 161 patients, where 51 were diagnosed as malignant, 64 as benign and 207 as normal. The
abnormalities are classified into microcalcification, circumscribed mass, ill-defined mass, spiculated mass,
architectural distortion, and asymmetry. In this study 322 mammogram images were selected as described
in Table 1. The original mammograms are 1024 x 1024 pixels, and almost 50% of the image comprised of
the background with a lot of noise. Therefore a cropping operation is applied to the images to cut off the
unwanted portions of the images. Regions of Interest (ROI) of 128x128 pixels are cropped manually, where
the given center of the abnormality area is selected to be the center of ROI. In the proposed CAD system,
microcalcifications are not considered, only circumscribed mass, ill-defined mass, spiculated mass,
architectural distortion, and asymmetry are considered. The preprocessing phase of digital mammograms
refers to the enhancement of mammograms intensity and contrast manipulation, noise reduction,
background removal, edges sharpening, filtering, etc.
2.2 ROI Image Enhancement
Histogram equalization [7] has been used for the reassignment of pixels to make the image better, as
proposed by H.D.Cheng et al [8]. In general, low contrast of mammographic images, hard to read masses in
mammograms [9], variation of the intensities of the masses such that radiopaque mass with high density
and radiolucent mass with low density in comparison with the background were found to be the reasons for
the enhancement [10]. Selective median filtering with CLAHE (Contrast Limited Adaptive Histogram
Equalization) algorithm will remove the noise and enhances the mammogram for better segmentation.
2.3 Segmentation
In analyzing mammogram image, it is important to distinguish the suspicious region from its surroundings.
The goal of segmentation is to simplify and 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 segments that
38](https://image.slidesharecdn.com/11-37-46segmentationandfeatureextractionoftumorsfromdigitalmammograms-120512235750-phpapp02/85/11-37-46-segmentation-and-feature-extraction-of-tumors-from-digital-mammograms-2-320.jpg)
![Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.4, 2012
collectively cover the entire image, or a set of contours extracted from the image [11].
Segmentation can be carried out using any of the standard techniques like Local Thresholding, Global
Thresholding, Region Growing, Region Clustering, Otsu Segmentation Technique, Template Matching. In
the proposed system, we have used Local Thresholding Technique [12 -14] and Otsu method [15] for
segmentation.
Local thresholding technique has been proven to provide an easy and convenient way to perform the
segmentation on digital mammogram. The segmentation is determined by a single value known as the
intensity threshold value. Then, each pixel in the image is compared with the threshold value. Pixel
intensity values higher than the threshold will result in a white spot in the output image
The second technique used for segmentation is Otsu’s method, which has shown a more satisfactory
performance in the medical image segmentation. It has been found to perform well compared to other
thresholding methods in segmenting the masses in digital mammogram.
2.4 Filter
After the segmentation process, there are still noises in the image which will affect the next stage, Feature
Extraction stage. These noises have to be suppressed using any of the filtering techniques that are
appropriate. Usually noises occupy less number of pixels compared to the tumor. These noises can be easily
be visualized by observing the segmentation output and they look like small white spots. The observed
pixels which are to be noises are removed by filling their intensity values by zeros.
2.5 Feature Extraction
Feature is used to denote a piece of information which is relevant for solving the computational
task related to a certain application. More specifically, features can refer to:
· The result of a general neighborhood operation (feature extractor or feature detector) applied to the
image,
• Specific structures in the image itself, ranging from simple structures such as points or edges to
more complex structures such as objects.
Feature extraction is the important step in breast cancer detection. Texture feature is important for image
classification. Various techniques have been used for computing texture features [16]
The recent use of textural features and machine learning (ML) classifiers has established a new research
direction to detect breast cancer. Texture features have been widely used as classification of masses in
digital mammogram. Texture features can be classified into three classes based on what they are derived
from Gray Level Co-Occurrence Matrices (GLCM), Gray Level Difference Statistics (GLDS), and Run
Length Statistics (RLS) [17]. Many studies have been focused on statistical texture features [18] used the
three types of statistical texture feature for masses detection and classification, [19] used the RLS to extract
the texture information from the region of interest. [20] Texture features in detection of masses in
mammogram. [21] Used texture analysis for the classification of mammographic masses. [22] Used
statistical texture features for early detection of masses in digitized mammogram.
Texture features is considered to be one of the widely used tool in masses detection in digital mammogram.
The Haralick features and co-occurrence matrix are the most methods of textural analysis for feature
extraction which are used to classify the regions of interest into benign and malignant masses. Most of the
texture analysis methods are applied directly on the original or filtering images. The computational of the
co-occurrence matrix to the whole image is required large computing time. However decreasing the size of
the co-occurrence matrix by reducing the number of gray levels in the image without losing the
information’s is must.
Statistical texture features have been proven to be powerful in classifying masses and normal breast tissues
39](https://image.slidesharecdn.com/11-37-46segmentationandfeatureextractionoftumorsfromdigitalmammograms-120512235750-phpapp02/85/11-37-46-segmentation-and-feature-extraction-of-tumors-from-digital-mammograms-3-320.jpg)
![Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.4, 2012
[23]. The implementation of feature extraction procedure relies on the quality of the texture, which is the
main descriptor for all the mammograms.
In the proposed system, Co-occurrence matrix is used for feature extraction. GLCM is a powerful tool for
image feature extraction. Gray level pixel distribution described by statistics like probability of two pixels
having particular gray level at particular spatial relationships. This spatial information is provided as two
dimensional gray level matrices. GLCM is a powerful tool for image feature extraction. Gray level pixel
distribution described by statistics like probability of two pixels having particular gray level at particular
spatial relationships. This spatial information is provided as two dimensional gray level matrices. Various
statistical measures such as correlation, energy, entropy, homogeneity and other statistical measures are
calculated based on GLCM for the output of fourth stage. The size of GLCM is determined by number of
gray level in an image.
The standard Haralick GLCM texture descriptors from Haralick (1973), GLCM Texture Descriptors from
Clausi (2002), GLCM Texture Descriptors from Soh and Tsatsoulis (1999) and MATLAB Image Processing
Toolbox (IPT) GLCM Texture descriptors are depicted in table 2, table 3, table 4 and table 5
respectively. In the proposed system, combination of all the four texture descriptors and statistical features
are extracted.
Apart from the texture features other statistical features like Mean, Standard Deviation, Variance,
Skewness, Kurtosis, Root Mean Square is determined.
3. EXPERIMENTAL OUTPUT
MATLAB has been used to calculate all the features required and GUIDE toolbox is used to develop the
GUI for the proposed system. In the first step, in the input image the ROI is selected by manual cropping
process. For the cropped image, image enhancement techniques have been employed to improve the image
quality. The segmentation is carried out for the enhanced cropped image by using Local thresholding
and Otsu segmentation techniques. Before feature extraction of this segmented image, noise is removed by
filling their intensity values by zero. These noises will appear as tiny white spots in the image. Features are
calculated for the noise suppressed segmented output. The figure 3 depicts the steps carried out.
4. CONCLUSION AND FUTURE SCOPE
It has been observed that after the segmentation process, then also white spots-Noise will appear which will
affects the feature extraction process. This noise has to be removed if feature extraction process has to
yield the accurate results. Once features are calculated for the entire MIAS database, then feature set has to
be constructed. The feature set has to be constructed in such a manner that it has to be understandable for
the classifiers. Any of the classifiers like Artificial Neural Network, Support Vector Machines, and Decision
Trees can be used for the classification of breast tumors. The accuracy of the classifier has to be
determined. The accuracy of any classifier will depend on the features extraction phase. The classifier has
to be trained for the feature set constructed. Then it has to be tested with unknown samples. The accuracy is
calculated for training and testing phases. For the training and testing the classifiers, the MIAS dataset can
be used. The classifiers can also be experimented for the digital mammograms collected from the local
oncology hospitals.
References
[1] L. Tabar and P. B. Dean. Mammography and breast cancer: the new era. Gynaecol Obstet, 82:319–326,
2003.
[2] M. L. Giger, N. Karssemeijer and S. G. Armato, “Computer-aided diagnosis in medical imaging”, IEEE
40](https://image.slidesharecdn.com/11-37-46segmentationandfeatureextractionoftumorsfromdigitalmammograms-120512235750-phpapp02/85/11-37-46-segmentation-and-feature-extraction-of-tumors-from-digital-mammograms-4-320.jpg)
![Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.4, 2012
Trans. on Med. Imaging, vol. 20, pp. 1205-1208, 2001.
[3] M. L. Giger, “Computer-aided diagnosis of breast lesions in medical images”, Comput. Science
Engineering, vol. 2, pp. 39-45, 2000.
[4] K. Doi, H. MacMahon, S. Katsuragawa, R. M. Nishikawa and Y. Jiang, “Computer-aided diagnosis in
radiology: potential and pitfalls”, Eur. J. Radiol., vol. 31, pp. 97-109, 1999.
[5] C. J. Vyborny, M. L. Giger and R. M. Nishikawa, “Computer-aided detection and diagnosis of breast
cancer”, Radiologic Clinics of North America, vol. 38, pp. 725-740, 2000
[6] http://peipa.essex.ac.uk/ipa/pix/mias.
[7] S.M.Pizer, E.O.P.Amburn and J.D.Austin, 1987. "Adaptive histogram equalization and its
variations”, Comput. Vision Graphics Image Process. 39, pp. 355-368.
[8] H.D.Cheng, X.J.Shi, R.Min, L.M.Hu, X.P.Cai and H.N.Du, 2006. "Approaches for Automated
Detection and Classification of Masses in Mammograms", Pattern Recognition 39, pp. 646- 668.
[9] Jacob Scharcanski and Claudio Rosito Jung, 2006. “Denoising and enhancing digital mammographic
images for visual screening”, Computerized Medical Imaging and Graphics 30, pp. 243–254.
[10] R.M. Rangayyan, L. Shen, Y. Shen, J.E.L. Desautels, H. Bryant, T.J. Terry, N. Horeczko and M.S.
Rose, 1997. “Improvement of sensitivity of breast cancer diagnosis with adaptive neighbourhood contrast
enhancement of mammograms”, IEEE Trans. Inform. Technol.Biomed. 1, pp. 161-170.
[11] G.M. Brake and N. Karssemeijer, 2001. “Segmentation of suspicious densities”, Med. Phys. 28, pp.
258-266.
[12] N. Karssemeijer. Automated classification of parenchymal patterns in mammograms. Phys. Med.
Biol., 43:365–378, 1998.
[13] Sung-Nien Yu, Kuan-Yuei Li and Yu-Kun Haung, 2006. "Detection of microcalcifications in digital
mammograms using wavelet filter and Markov random field model", Computerized Medical Imaging and
Graphics 30, pp. 163-173.
[14] Xu Kai, Qin Kun and Pei Tao, 2006. “Interactive Method for Image Segmentation Based on Cloud
Model”, Computer Engineering and Application 34, pp. 33-35.
[15] Jeong HJ, Kim TY, Hwang HG et al. Comparison of thresholding methods for breast tumor cell
segmentation. IEEE Proceedings of 7th International Workshop on Enterprise Networking and Computing
in Healthcare Industry. 2005: 392-5.
[16] H.S.Sheshadri and A.Kandaswamy,”Experimental investigation on breast tissue classification based
on statistical feature extraction of mammograms”, Computerized Medical Imaging and Graphics, no.31,
pp.46-48, 2007.
[17] H.D. Cheng, X.J. Shi, R. Min, L.M. Hu, X.P. Cai, H.N. Du (2006) “Approaches for automated
detection and classification of masses in mammograms”, Pattern Recognition, Vol. 39, pp. 646-668.
[18] Mavroforakis,M.E.,Georgiou,H.V.,Dimitropoulos,N., Cavouras,D., and Theodoridis,S.,
“Mammographic masses characterization based on localized texture and dataset fractal analysis using
linear, neural and support vector
machine classifiers”, Artificial Intelligence in Medicine, 2006, pp. 145—162.
[19]Szekely, N, Toth, N. Pataki, B. (2004). A hybrid system for detecting masses in mammographic
images. Instrumentation and measurement technology conference, 2004. IMTC 04. Proceeding of the 21st
IEEE Vol3, 18-20 May 2004 pp2065-2070.
[20] K. Bovis and S. Singh. Detection of masses in mammograms using texture features. 15th International
Conference on Pattern Recognition (ICPR'00), 2:2267, 2000.
[21] Mudigonda, N.R, Rangayyan, R. Desautels, J.E.L (2000); Gradients and texture analysis for the
classification of mammographic masses Medical Image, IEEE Transaction on Vol 19, Issue 10, Oct. 2000
41](https://image.slidesharecdn.com/11-37-46segmentationandfeatureextractionoftumorsfromdigitalmammograms-120512235750-phpapp02/85/11-37-46-segmentation-and-feature-extraction-of-tumors-from-digital-mammograms-5-320.jpg)
![Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol 3, No.4, 2012
pp 1032-1043.
[22] N. Youssry, F.E.Z. Abou-Chadi, and A.M. El-Sayad. Early detection of masses in digitized
mammograms using texture features and neuro-fuzzy model. 4th Annual IEEE Conf on Information
Technology Applications in
Biomedicine, 2003.
[23] Haralick,R.M., Shanmugam,K., Dinstein,I., “Textural features for image classification”, IEEE Trans Sys
ManCyb, 1973, pp. 610—21.
Figure 1. Sample mammogram
Figure 2. Proposed CAD System
42](https://image.slidesharecdn.com/11-37-46segmentationandfeatureextractionoftumorsfromdigitalmammograms-120512235750-phpapp02/85/11-37-46-segmentation-and-feature-extraction-of-tumors-from-digital-mammograms-6-320.jpg)
11.[37 46]segmentation and feature extraction of tumors from digital mammograms
- 1. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.4, 2012 Segmentation and Feature Extraction of Tumors from Digital Mammograms PRADEEP N *, Department of Computer Science and Engineering, Bapuji Institute of Engineering and Technology, Davangere, Karnataka, INDIA. Email: nmnpradeep@yahoo.com GIRISHA H, Department of Computer Science and Engineering, Rao Bahadur Y Mahabaleswarappa Engineering College, Bellary, Karnataka, INDIA. KARIBASAPPA K, Department of Computer Science and Engineering, Dayananda Sagar College of Engineering, Bangalore, Karnataka, INDIA. Abstract Mammography is one of the available techniques for the early detection of masses or abnormalities which is related to breast cancer. Breast Cancer is the uncontrolled of cells in the breast region, which may affect the other parts of the body. The most common abnormalities that might indicate breast cancer are masses and calcifications. Masses appear in a mammogram as fine, granular clusters and also masses will not have sharp boundaries, so often difficult to identify in a raw mammogram. Digital Mammography is one of the best available technologies currently being used for the early detection of breast cancer. Computer Aided Detection System has to be developed for the detection of masses and calcifications in Digital Mammogram, which acts as a secondary tool for the radiologists for diagnosing the breast cancer. In this paper, we have proposed a secondary tool for the radiologists that help them in the segmentation and feature extraction process. Keywords: Mammography, Breast Cancer, Masses, Calcification, Digital Mammography, Computer Aided Detection System, Segmentation, Feature Extraction 1. INTRODUCTION Cancer is an abnormal, continual multiplying of cells. The cells divide uncontrollably and may grow into adjacent tissue or spread to distant parts of the body. Breast cancer remains a leading cause of cancer deaths among women in many parts of the world. Early detection of breast cancer through periodic screening has noticeably improved the outcome of the disease [1]. The mass of cancer cells will eventually become large enough to produce lumps, masses, or tumors that can be detected. Tumor is uncontrolled growth of cells which can be either Benign or Malignant. Benign Tumors are not cancerous. Benign tumors may grow larger but do not spread to other parts of the body. Malignant Tumors is cancerous. Malignant tumors can invade and destroy nearby tissue and spread to other parts of the body. Tumor can be easily identified in mammogram because tumor part is highly bright (having high intensity) compared to other part 37
- 2. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.4, 2012 (background) of the mammogram image as shown in Figure 1. In the figure 1, we can observe that the marked oval shape area have higher intensity compared to the surrounding area. This marked oval shape is the required Region Of Interest (ROI). Segmentation techniques can be used to extract the tumor from the mammogram. Computer Aided (CA) detection systems have been developed to aid radiologists in detecting mammographic abnormalities [2-5]. There are large numbers of diagnostic methods currently available for the diagnosis of Breast Cancer, among which digital mammography is the most reliable method, for detecting early breast cancer [6-7]. Early detection of Breast Cancer can decrease the mortality rate. In this paper, we have proposed a CAD system to segment the tumor from the ROI and calculate the features that have importance in the classification of tumors. 2. PROPOSED CAD SYSTEM The proposed CAD system consists of five principal stages. The first stage is the collection of input images and ROI extraction, second stage is ROI Image Enhancement, the third stage is Segmentation, the fourth stage is Filter and the last stage is Feature Enhancement. 2.1 ROI Extraction The input images are collected from Mammographic Image Analysis Society (MIAS) research database. The MIAS dataset [6] is used to test the proposed technique. These images were previously investigated and labeled by an expert radiologist based on technical experience and biopsy. The dataset is selected due to the various cases it includes. The dataset is composed of 322 mammograms of right and left breast, from 161 patients, where 51 were diagnosed as malignant, 64 as benign and 207 as normal. The abnormalities are classified into microcalcification, circumscribed mass, ill-defined mass, spiculated mass, architectural distortion, and asymmetry. In this study 322 mammogram images were selected as described in Table 1. The original mammograms are 1024 x 1024 pixels, and almost 50% of the image comprised of the background with a lot of noise. Therefore a cropping operation is applied to the images to cut off the unwanted portions of the images. Regions of Interest (ROI) of 128x128 pixels are cropped manually, where the given center of the abnormality area is selected to be the center of ROI. In the proposed CAD system, microcalcifications are not considered, only circumscribed mass, ill-defined mass, spiculated mass, architectural distortion, and asymmetry are considered. The preprocessing phase of digital mammograms refers to the enhancement of mammograms intensity and contrast manipulation, noise reduction, background removal, edges sharpening, filtering, etc. 2.2 ROI Image Enhancement Histogram equalization [7] has been used for the reassignment of pixels to make the image better, as proposed by H.D.Cheng et al [8]. In general, low contrast of mammographic images, hard to read masses in mammograms [9], variation of the intensities of the masses such that radiopaque mass with high density and radiolucent mass with low density in comparison with the background were found to be the reasons for the enhancement [10]. Selective median filtering with CLAHE (Contrast Limited Adaptive Histogram Equalization) algorithm will remove the noise and enhances the mammogram for better segmentation. 2.3 Segmentation In analyzing mammogram image, it is important to distinguish the suspicious region from its surroundings. The goal of segmentation is to simplify and 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 segments that 38
- 3. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.4, 2012 collectively cover the entire image, or a set of contours extracted from the image [11]. Segmentation can be carried out using any of the standard techniques like Local Thresholding, Global Thresholding, Region Growing, Region Clustering, Otsu Segmentation Technique, Template Matching. In the proposed system, we have used Local Thresholding Technique [12 -14] and Otsu method [15] for segmentation. Local thresholding technique has been proven to provide an easy and convenient way to perform the segmentation on digital mammogram. The segmentation is determined by a single value known as the intensity threshold value. Then, each pixel in the image is compared with the threshold value. Pixel intensity values higher than the threshold will result in a white spot in the output image The second technique used for segmentation is Otsu’s method, which has shown a more satisfactory performance in the medical image segmentation. It has been found to perform well compared to other thresholding methods in segmenting the masses in digital mammogram. 2.4 Filter After the segmentation process, there are still noises in the image which will affect the next stage, Feature Extraction stage. These noises have to be suppressed using any of the filtering techniques that are appropriate. Usually noises occupy less number of pixels compared to the tumor. These noises can be easily be visualized by observing the segmentation output and they look like small white spots. The observed pixels which are to be noises are removed by filling their intensity values by zeros. 2.5 Feature Extraction Feature is used to denote a piece of information which is relevant for solving the computational task related to a certain application. More specifically, features can refer to: · The result of a general neighborhood operation (feature extractor or feature detector) applied to the image, • Specific structures in the image itself, ranging from simple structures such as points or edges to more complex structures such as objects. Feature extraction is the important step in breast cancer detection. Texture feature is important for image classification. Various techniques have been used for computing texture features [16] The recent use of textural features and machine learning (ML) classifiers has established a new research direction to detect breast cancer. Texture features have been widely used as classification of masses in digital mammogram. Texture features can be classified into three classes based on what they are derived from Gray Level Co-Occurrence Matrices (GLCM), Gray Level Difference Statistics (GLDS), and Run Length Statistics (RLS) [17]. Many studies have been focused on statistical texture features [18] used the three types of statistical texture feature for masses detection and classification, [19] used the RLS to extract the texture information from the region of interest. [20] Texture features in detection of masses in mammogram. [21] Used texture analysis for the classification of mammographic masses. [22] Used statistical texture features for early detection of masses in digitized mammogram. Texture features is considered to be one of the widely used tool in masses detection in digital mammogram. The Haralick features and co-occurrence matrix are the most methods of textural analysis for feature extraction which are used to classify the regions of interest into benign and malignant masses. Most of the texture analysis methods are applied directly on the original or filtering images. The computational of the co-occurrence matrix to the whole image is required large computing time. However decreasing the size of the co-occurrence matrix by reducing the number of gray levels in the image without losing the information’s is must. Statistical texture features have been proven to be powerful in classifying masses and normal breast tissues 39
- 4. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.4, 2012 [23]. The implementation of feature extraction procedure relies on the quality of the texture, which is the main descriptor for all the mammograms. In the proposed system, Co-occurrence matrix is used for feature extraction. GLCM is a powerful tool for image feature extraction. Gray level pixel distribution described by statistics like probability of two pixels having particular gray level at particular spatial relationships. This spatial information is provided as two dimensional gray level matrices. GLCM is a powerful tool for image feature extraction. Gray level pixel distribution described by statistics like probability of two pixels having particular gray level at particular spatial relationships. This spatial information is provided as two dimensional gray level matrices. Various statistical measures such as correlation, energy, entropy, homogeneity and other statistical measures are calculated based on GLCM for the output of fourth stage. The size of GLCM is determined by number of gray level in an image. The standard Haralick GLCM texture descriptors from Haralick (1973), GLCM Texture Descriptors from Clausi (2002), GLCM Texture Descriptors from Soh and Tsatsoulis (1999) and MATLAB Image Processing Toolbox (IPT) GLCM Texture descriptors are depicted in table 2, table 3, table 4 and table 5 respectively. In the proposed system, combination of all the four texture descriptors and statistical features are extracted. Apart from the texture features other statistical features like Mean, Standard Deviation, Variance, Skewness, Kurtosis, Root Mean Square is determined. 3. EXPERIMENTAL OUTPUT MATLAB has been used to calculate all the features required and GUIDE toolbox is used to develop the GUI for the proposed system. In the first step, in the input image the ROI is selected by manual cropping process. For the cropped image, image enhancement techniques have been employed to improve the image quality. The segmentation is carried out for the enhanced cropped image by using Local thresholding and Otsu segmentation techniques. Before feature extraction of this segmented image, noise is removed by filling their intensity values by zero. These noises will appear as tiny white spots in the image. Features are calculated for the noise suppressed segmented output. The figure 3 depicts the steps carried out. 4. CONCLUSION AND FUTURE SCOPE It has been observed that after the segmentation process, then also white spots-Noise will appear which will affects the feature extraction process. This noise has to be removed if feature extraction process has to yield the accurate results. Once features are calculated for the entire MIAS database, then feature set has to be constructed. The feature set has to be constructed in such a manner that it has to be understandable for the classifiers. Any of the classifiers like Artificial Neural Network, Support Vector Machines, and Decision Trees can be used for the classification of breast tumors. The accuracy of the classifier has to be determined. The accuracy of any classifier will depend on the features extraction phase. The classifier has to be trained for the feature set constructed. Then it has to be tested with unknown samples. The accuracy is calculated for training and testing phases. For the training and testing the classifiers, the MIAS dataset can be used. The classifiers can also be experimented for the digital mammograms collected from the local oncology hospitals. References [1] L. Tabar and P. B. Dean. Mammography and breast cancer: the new era. Gynaecol Obstet, 82:319–326, 2003. [2] M. L. Giger, N. Karssemeijer and S. G. Armato, “Computer-aided diagnosis in medical imaging”, IEEE 40
- 5. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.4, 2012 Trans. on Med. Imaging, vol. 20, pp. 1205-1208, 2001. [3] M. L. Giger, “Computer-aided diagnosis of breast lesions in medical images”, Comput. Science Engineering, vol. 2, pp. 39-45, 2000. [4] K. Doi, H. MacMahon, S. Katsuragawa, R. M. Nishikawa and Y. Jiang, “Computer-aided diagnosis in radiology: potential and pitfalls”, Eur. J. Radiol., vol. 31, pp. 97-109, 1999. [5] C. J. Vyborny, M. L. Giger and R. M. Nishikawa, “Computer-aided detection and diagnosis of breast cancer”, Radiologic Clinics of North America, vol. 38, pp. 725-740, 2000 [6] http://peipa.essex.ac.uk/ipa/pix/mias. [7] S.M.Pizer, E.O.P.Amburn and J.D.Austin, 1987. "Adaptive histogram equalization and its variations”, Comput. Vision Graphics Image Process. 39, pp. 355-368. [8] H.D.Cheng, X.J.Shi, R.Min, L.M.Hu, X.P.Cai and H.N.Du, 2006. "Approaches for Automated Detection and Classification of Masses in Mammograms", Pattern Recognition 39, pp. 646- 668. [9] Jacob Scharcanski and Claudio Rosito Jung, 2006. “Denoising and enhancing digital mammographic images for visual screening”, Computerized Medical Imaging and Graphics 30, pp. 243–254. [10] R.M. Rangayyan, L. Shen, Y. Shen, J.E.L. Desautels, H. Bryant, T.J. Terry, N. Horeczko and M.S. Rose, 1997. “Improvement of sensitivity of breast cancer diagnosis with adaptive neighbourhood contrast enhancement of mammograms”, IEEE Trans. Inform. Technol.Biomed. 1, pp. 161-170. [11] G.M. Brake and N. Karssemeijer, 2001. “Segmentation of suspicious densities”, Med. Phys. 28, pp. 258-266. [12] N. Karssemeijer. Automated classification of parenchymal patterns in mammograms. Phys. Med. Biol., 43:365–378, 1998. [13] Sung-Nien Yu, Kuan-Yuei Li and Yu-Kun Haung, 2006. "Detection of microcalcifications in digital mammograms using wavelet filter and Markov random field model", Computerized Medical Imaging and Graphics 30, pp. 163-173. [14] Xu Kai, Qin Kun and Pei Tao, 2006. “Interactive Method for Image Segmentation Based on Cloud Model”, Computer Engineering and Application 34, pp. 33-35. [15] Jeong HJ, Kim TY, Hwang HG et al. Comparison of thresholding methods for breast tumor cell segmentation. IEEE Proceedings of 7th International Workshop on Enterprise Networking and Computing in Healthcare Industry. 2005: 392-5. [16] H.S.Sheshadri and A.Kandaswamy,”Experimental investigation on breast tissue classification based on statistical feature extraction of mammograms”, Computerized Medical Imaging and Graphics, no.31, pp.46-48, 2007. [17] H.D. Cheng, X.J. Shi, R. Min, L.M. Hu, X.P. Cai, H.N. Du (2006) “Approaches for automated detection and classification of masses in mammograms”, Pattern Recognition, Vol. 39, pp. 646-668. [18] Mavroforakis,M.E.,Georgiou,H.V.,Dimitropoulos,N., Cavouras,D., and Theodoridis,S., “Mammographic masses characterization based on localized texture and dataset fractal analysis using linear, neural and support vector machine classifiers”, Artificial Intelligence in Medicine, 2006, pp. 145—162. [19]Szekely, N, Toth, N. Pataki, B. (2004). A hybrid system for detecting masses in mammographic images. Instrumentation and measurement technology conference, 2004. IMTC 04. Proceeding of the 21st IEEE Vol3, 18-20 May 2004 pp2065-2070. [20] K. Bovis and S. Singh. Detection of masses in mammograms using texture features. 15th International Conference on Pattern Recognition (ICPR'00), 2:2267, 2000. [21] Mudigonda, N.R, Rangayyan, R. Desautels, J.E.L (2000); Gradients and texture analysis for the classification of mammographic masses Medical Image, IEEE Transaction on Vol 19, Issue 10, Oct. 2000 41
- 6. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.4, 2012 pp 1032-1043. [22] N. Youssry, F.E.Z. Abou-Chadi, and A.M. El-Sayad. Early detection of masses in digitized mammograms using texture features and neuro-fuzzy model. 4th Annual IEEE Conf on Information Technology Applications in Biomedicine, 2003. [23] Haralick,R.M., Shanmugam,K., Dinstein,I., “Textural features for image classification”, IEEE Trans Sys ManCyb, 1973, pp. 610—21. Figure 1. Sample mammogram Figure 2. Proposed CAD System 42
- 7. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.4, 2012 Class Benign Malignant Total Microcalcification 12 13 25 Circumscribed 19 4 23 ll-defined 7 7 14 Spiculated 11 8 19 Architectural 9 10 19 Asymmetry 6 9 15 Normal tissue 9 - 207 Total 64 51 322 Table 1. The Distribution of MIAS Data Set 43
- 8. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.4, 2012 Table 2: Standard GLCM Texture Descriptors (Haralick, 1973) 44
- 9. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.4, 2012 45
- 10. Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.4, 2012 46
- 11. International Journals Call for Paper The IISTE, a U.S. publisher, is currently hosting the academic journals listed below. The peer review process of the following journals usually takes LESS THAN 14 business days and IISTE usually publishes a qualified article within 30 days. Authors should send their full paper to the following email address. More information can be found in the IISTE website : www.iiste.org Business, Economics, Finance and Management PAPER SUBMISSION EMAIL European Journal of Business and Management EJBM@iiste.org Research Journal of Finance and Accounting RJFA@iiste.org Journal of Economics and Sustainable Development JESD@iiste.org Information and Knowledge Management IKM@iiste.org Developing Country Studies DCS@iiste.org Industrial Engineering Letters IEL@iiste.org Physical Sciences, Mathematics and Chemistry PAPER SUBMISSION EMAIL Journal of Natural Sciences Research JNSR@iiste.org Chemistry and Materials Research CMR@iiste.org Mathematical Theory and Modeling MTM@iiste.org Advances in Physics Theories and Applications APTA@iiste.org Chemical and Process Engineering Research CPER@iiste.org Engineering, Technology and Systems PAPER SUBMISSION EMAIL Computer Engineering and Intelligent Systems CEIS@iiste.org Innovative Systems Design and Engineering ISDE@iiste.org Journal of Energy Technologies and Policy JETP@iiste.org Information and Knowledge Management IKM@iiste.org Control Theory and Informatics CTI@iiste.org Journal of Information Engineering and Applications JIEA@iiste.org Industrial Engineering Letters IEL@iiste.org Network and Complex Systems NCS@iiste.org Environment, Civil, Materials Sciences PAPER SUBMISSION EMAIL Journal of Environment and Earth Science JEES@iiste.org Civil and Environmental Research CER@iiste.org Journal of Natural Sciences Research JNSR@iiste.org Civil and Environmental Research CER@iiste.org Life Science, Food and Medical Sciences PAPER SUBMISSION EMAIL Journal of Natural Sciences Research JNSR@iiste.org Journal of Biology, Agriculture and Healthcare JBAH@iiste.org Food Science and Quality Management FSQM@iiste.org Chemistry and Materials Research CMR@iiste.org Education, and other Social Sciences PAPER SUBMISSION EMAIL Journal of Education and Practice JEP@iiste.org Journal of Law, Policy and Globalization JLPG@iiste.org Global knowledge sharing: New Media and Mass Communication NMMC@iiste.org EBSCO, Index Copernicus, Ulrich's Journal of Energy Technologies and Policy JETP@iiste.org Periodicals Directory, JournalTOCS, PKP Historical Research Letter HRL@iiste.org Open Archives Harvester, Bielefeld Academic Search Engine, Elektronische Public Policy and Administration Research PPAR@iiste.org Zeitschriftenbibliothek EZB, Open J-Gate, International Affairs and Global Strategy IAGS@iiste.org OCLC WorldCat, Universe Digtial Library , Research on Humanities and Social Sciences RHSS@iiste.org NewJour, Google Scholar. Developing Country Studies DCS@iiste.org IISTE is member of CrossRef. All journals Arts and Design Studies ADS@iiste.org have high IC Impact Factor Values (ICV).