Ali El-Zaart

A Comparison of SVM Kernel Functions for Breast Cancer Detection

Support vector machines outperform other classification methods for breast cancer detection. However the performance of SVM is greatly affected by the choice of a kernel function among other factors. This article presents a comparative study of different kernel functions for breast cancer detection. The focus is on classification using SVM with different kernel functions. The comparison with neural network based method using MLP is also given. Furthermore, we examine the affect of selecting feature subsets before applying classification with different kernels. For features subset selection we used genetic algorithm. The evaluation is based on 5 X 2 cross validation.

Images thresholding using ISODATA technique with gamma distribution

Image segmentation is a fundamental step in many applications of image processing. Many image segmentation techniques exist based on different methods such as classification-based methods, edge-based methods, region-based methods, and hybrid methods. The principal approach of segmentation is based on thresholding (classification) that is related to thresholds estimation problem. The ISODATA (Iterative Self-Organizing Data Analysis Technique) method is one of the classification-based methods in image segmentation. We assumed that the data in images is modeled by Gamma distribution. The objective of this paper is to explain a new method that combines Gamma distribution with the technique of ISODATA. The algorithm has two phases: splitting using Gamma distribution then merging which are done based on some predefined parameters. Experimental results showed good segmentation for artificial and real images.

Feature extraction values for breast cancer mammography images

Breast cancer is one of the most common cancers among woman of the developing countries in the world, and it has also become a major cause of death. Treatment of breast cancer is effective only if it is detected at an early stage. X-ray mammography is the most effective method for early detection but the mammography images are complex. Thus nowadays, image processing and image analysis techniques are use to assist radiologist for detecting tumors in mammography images. In this paper we specify and determined the important and significant Breast Cancer Feature Extraction. After that, we analyze Breast cancer mammography images using these significant features. However, the aim of this study is to determine the features extraction ranges values for Breast Cancer mammography image.

Expectation-maximization technique for fibro-glandular discs detection in mammography images

Breast cancer is among the leading causes of death in women worldwide. Mammography is the most effective imaging method for detecting no-palpable early-stage breast cancer. Understanding the nature of data in mammography images is very important for developing a model that fits well the data. Statistical distributions are widely used on the modelling of the data. Gamma distribution is more suitable than Gaussian distribution for modelling the data in mammography images. In this paper, we will use Gamma distribution to model the data in mammography images. The histogram of images can be seen as a mixture of Gamma distributions. Thresholds are selected at the valleys of a multi-modal histogram. The estimation of thresholds is based on the statistical parameters of the histogram. The expectation-maximization technique with gamma distribution (EMTG) is therefore developed to estimate the statistical histogram parameters. The experimental results on mammography images using this technique showed improvement in the accuracy in detection of the fibro-glandular discs.

Finite Gamma mixture modelling using minimum message length inference: application to SAR image analysis

This paper discusses the unsupervised learning problem for finite mixtures of Gamma distributions. An important part of this problem is determining the number of clusters which best describes a set of data. We apply the Minimum Message Length (MML) criterion to the unsupervised learning problem in the case of finite mixtures of Gamma distributions. The MML and other criteria in the literature are compared in terms of their ability to estimate the number of clusters in a data set. The comparison utilizes synthetic and RADARSAT SAR images. The performance of our method is also tested by contextual evaluations involving SAR image segmentation and change detection.

Minimum Cross Entropy Thresholding for SAR Images

Vision plays the most important role in human perception, which is limited to only the visual band of the electromagnetic spectrum. Therefore, the need for Radar imaging systems, to recover some sources that are not within human visual band, is raised. This paper presents a new algorithm for Synthetic Aperture Radar (SAR) images segmentation based on thresholding technique. Generally, segmentation of a SAR image falls into two categories; one based on grey levels and the other based on texture. The present paper deals with SAR images segmentation based on grey levels. We developed a new formula using Minimum Cross Entropy Thresholding (MCET) method for estimating optimal threshold value based on Gamma distribution to analyzing data on images; that means histogram of SAR images is assumed to be a mixture of Gamma distributions. The proposed method is iterative which decreases the number of operation to converge tends to the optimal solution. It is applied on bi-modal and multimodal scenarios. The results obtained are promising.

PAVING THE WAY TO SMART SUSTAINABLE CITIES: TRANSFORMATION MODELS AND CHALLENGES

Rapid urbanization and globalization make the move toward Smart Sustainable Cities (SSC) a must. Achieving successful transformation towards SSCs constitutes a significant challenge for policy makers. One area that is not well covered in the literature is the application of SSCs in specific regions, such as the Arab region. This paper draws upon examples of SSCs initiatives and existing SSC transformation frameworks to more fully articulate the challenges of achieving successful SSC projects across the Arab region. One of the interesting emergent themes is the emergence of two main approaches to SSCs transformation, Brownfield and Greenfield approaches.

A novel approach for Braille images segmentation

Braille recognition is the ability to detect and recognize Braille characters embossed on Braille document. The result is used in several applications such as embossing, printing, translating...etc. However, the performance of these applications is affected by poor quality imaging due to several factors such as scanner quality, scan resolution, lighting, and type of embossed documents. In this work, we extend previous research efforts on Braille recognition systems by proposing a new method for Braille image segmentation using Between-Class Variance with Gamma distribution. The technique of Between-Class Variance was proposed by Otsu using a mixture of Gaussian distributions. Gaussian distribution is widely used for modeling the histogram of images, but due to the asymmetric nature of the distribution of gray levels in Braille images, Gamma distribution is more suitable. The proposed method is composed of two main parts. (a) Find the optimal estimated threshold values using Between-Class Variance with a mixture of Gamma distributions. (b) Use the optimal estimated thresholds values to segment Braille images. Our method was applied on several Braille images scanned by flatbed scanner. The experimental results on Braille images using this technique showed improvement in the accuracy of Braille images segmentation.

Contrast Enhancement of MRI Images

The technique of modification of the histogram of an image can be applied to the problem of image enhancement. Global histogram equalization and local area histogram equalization are two well-known techniques for the same purpose. In this paper new method is proposed to enhance the contrast of bimodal MRI images using histogram specification with Gamma distribution. The method is aimed to read the original image and calculate its histogram original histogram then apply the Maximum Likelihood Gamma Distribution (MLGD) method to get an accurate statistical information of the original histogram as the means and prior probabilities of the two modes, then we separate the two modes by shift the first mode left or shift the second mode right or perform both shifts. After that we will generate a new histogram called “Desired Histogram” using the new data. By applying a histogram specification method, a high contrast image will be produced. The new method of contrast enhancement of MRI image using histogram specification with Gamma distribution has been tested and showed good results.

A new approach for pupil detection in iris recognition system

Recently, Personal Identification System becomes a key factor for safety and secures environments. Iris segmentation step is one of the most important steps that plays a vital role in the accuracy and efficiency of Personal Identification System. Pupil detection and isolation is the foremost task in this step. Many researches had proposed several methods to detect the pupil boundary. Some of those methods detect the pupil by an estimated threshold or by search for the value of the first dark region, and consider it as an estimated threshold. But, the dark intensities of pupil are not fixed due to the coming lighting. Other iris segmentation methods which are used for iris recognition technology consider the boundary of pupil is a circle. But, the boundary of pupil is not quite circle and a small error in detecting this boundary will lead to lose some information surrounding the pupil. Another problem that makes most of the iris segmentation method fail to detect the pupil boundary is the head rotation or the eye rotation. As a result of these problems, the iris recognition system will be inefficient and inaccurate. In this paper a new method has developed to detect the boundary of pupil based on minimum and mean intensity of pupil. This method is able to detect and isolate pupil with high accuracy results approximately 100% and reduce time consuming