Neveen I. Ghali

Adaptive k-means clustering algorithm for MR breast image segmentation

Image segmentation is vital for meaningful analysis and interpretation of the medical images. The most popular method for clustering is k-means clustering. This article presents a new approach intended to provide more reliable magnetic resonance (MR) breast image segmentation that is based on adaptation to identify target objects through an optimization methodology that maintains the optimum result during iterations. The proposed approach improves and enhances the effectiveness and efficiency of the traditional k-means clustering algorithm. The performance of the presented approach was evaluated using various tests and different MR breast images. The experimental results demonstrate that the overall accuracy provided by the proposed adaptive k-means approach is superior to the standard k-means clustering technique.

Review article: Computational intelligence techniques in bioinformatics

Computational intelligence (CI) is a well-established paradigm with current systems having many of the characteristics of biological computers and capable of performing a variety of tasks that are difficult to do using conventional techniques. It is a methodology involving adaptive mechanisms and/or an ability to learn that facilitate intelligent behavior in complex and changing environments, such that the system is perceived to possess one or more attributes of reason, such as generalization, discovery, association and abstraction. The objective of this article is to present to the CI and bioinformatics research communities some of the state-of-the-art in CI applications to bioinformatics and motivate research in new trend-setting directions. In this article, we present an overview of the CI techniques in bioinformatics. We will show how CI techniques including neural networks, restricted Boltzmann machine, deep belief network, fuzzy logic, rough sets, evolutionary algorithms (EA), genetic algorithms (GA), swarm intelligence, artificial immune systems and support vector machines, could be successfully employed to tackle various problems such as gene expression clustering and classification, protein sequence classification, gene selection, DNA fragment assembly, multiple sequence alignment, and protein function prediction and its structure. We discuss some representative methods to provide inspiring examples to illustrate how CI can be utilized to address these problems and how bioinformatics data can be characterized by CI. Challenges to be addressed and future directions of research are also presented and an extensive bibliography is included.

Genetic Algorithms for community detection in social networks

Community detection in complex networks has attracted a lot of attention in recent years. Community detection can be viewed as an optimization problem, in which an objective function that captures the intuition of a community as a group of nodes with better internal connectivity than external connectivity is chosen to be optimized. Many single-objective optimization techniques have been used to solve the problem however those approaches have its drawbacks since they try optimizing one objective function and this results to a solution with a particular community structure property. More recently researchers viewed the problem as a multi-objective optimization problem and many approaches have been proposed to solve it. However which objective functions could be used with each other is still under debated since many objective functions have been proposed over the past years and in somehow most of them are similar in definition. In this paper we use Genetic Algorithm (GA) as an effective optimization technique to solve the community detection problem as a single-objective and multi-objective problem, we use the most popular objectives proposed over the past years, and we show how those objective correlate with each other, and their performances when they are used in the single-objective Genetic Algorithm and the Multi-Objective Genetic Algorithm and the community structure properties they tend to produce.

Evaluating the effects of image filters in CT Liver CAD system

The main objective of image pre-processing is to improve the quality of an image so that it makes subsequent phases of image analysis like segmentation or recognition easier or more effective. Filtering is a key pre-processing technique used for various effects including contrast stretching, sharpening and smoothing. In this paper, we evaluate and analyse the effect of several image filtering techniques with respect to their computer aided diagnosis (CAD) performance. The techniques we investigate include contrast stretching, convolution, median fitlering, averaging, inverse transformation and logarithm transformation filters. An application of CT liver imaging CAD was chosen and the selected filters were applied to see their ability and accuracy to segment and isolate the liver region of interest using a region growing segmentation approach. The effect of the filtering techniques on the segmentation performance of the CAD system was investigated using mean squared error (MSE) and similarity index (SI). The highest performance was achieved for a contrast stretching filter (MSE = 0.1869, SI = 0.8423) and the combination of contrast stretching and average filter (MSE = 0.17198 and SI = 0.83257).

Level set-based CT liver image segmentation with watershed and artificial neural networks

The objective of this paper is to evaluate a new combined approach intended for reliable CT liver image segmentation, to separate the liver from other organs, and segment the liver into a set of regions of interest (ROIs). The approach combines the level set with watershed approach used as post segmentation step to produce a reliable segmentation result. Features of first order statistics and grey-level cooccurrence matrix, are calculated and passed to an artificial neural network, to be trained and to classify infected regions. Filtering is used before the segmentation approach to enhance contrast, remove noise and emphasize certain features, as well as connecting ribs around the liver. To evaluate the performance of presented approach, we performed many tests on different CT liver images. The experimental results obtained, show that the overall accuracy offered by the proposed approach is 92.1% in segmenting CT liver images into set of regions even with noise, and 88.9% average accuracy for neural network classification.

A Hybrid Approach for Biometric Template Security

Privacy and Security has become an increasingly serious ýproblem for any biometric systems. Template protection, ýwhich mainly prevents from data loss and hacking the stored ýtemplates, is one of the most important issues when considering ýprivacy and security. Cancelable biometrics approach Scheme ýhas been proposed to address this problem. On the other hand ýthe symmetric hash functions might be used to increase the ýbiometric security level by making it is hard to attack the ýtemplate. In this paper, a new hybrid approach based on ýcombining approaches based on transformation and ýcryptosystem on fingerprints is applied. The results are ýcompared to a pre-presented approach in [1] showing that the ýnew approach is more efficient as the security of biometric ýtemplate is increased meanwhile the error rate is minimized.

Social Networks Analysis: Tools, Measures and Visualization

Social Network Analysis (SNA) is becoming an important tool for investigators, but all the necessary information is often available in a distributed environment. Currently there is no information system that helps managers and team leaders monitor the status of a social network. This chapter presents an overview of the basic concepts of social networks in data analysis including social network analysis metrics and performances. Different problems in social networks are discussed such as uncertainty, missing data and finding the shortest path in a social network. Community structure, detection and visualization in social network analysis is also illustrated. This chapter bridges the gap among the users by combining social network analysis methods and information visualization technology to help a user visually identify the occurrence of a possible relationship amongst the members in a social network. The chapter illustrates an online visualization method for a DBLP (Digital Bibliography Library Project) dataset of publications from the field of computer science, which is focused on the co-authorship relationship based on the intensity and topic of joint publications. Challenges to be addressed and future directions of research are presented and an extensive bibliography is also included.

Virtual Reality Technology for Blind and Visual Impaired People: Reviews and Recent Advances

Virtual reality technology enables people to become immersed in a computer-simulated and three-dimensional environment. In this chapter, we investigate the effects of the virtual reality technology on disabled people such as blind and visually impaired people (VIP) in order to enhance their computer skills and prepare them to make use of recent technology in their daily life. As well as, they need to advance their information technology skills beyond the basic computer training and skills. This chapter describes what best tools and practices in information technology to support disabled people such as deaf-blind and visual impaired people in their activities such as mobility systems, computer games, accessibility of e-learning, web-based information system, and wearable finger-braille interface for navigation of deaf-blind. Moreover, we will show how physical disabled people can benefits from the innovative virtual reality techniques and discuss some representative examples to illustrate how virtual reality technology can be utilized to address the information technology problem of blind and visual impaired people. Challenges to be addressed and an extensive bibliography are included.

Known-plaintext attack of DES-16 using Particle Swarm Optimization

Discovering the root key bits in the cryptanalysis of 16-rounded Data Encryption Standard (DES-160 is considered to be a hard problem. In this paper we present an approach for cryptanalysis of DES-16 based on Particle Swarm Optimization (PSO) using Known-plaintext attack and some equations that deduced from the relationship between sub-key differences and root key information. In Known-plaintext attack the cryptanalyst possesses one or more plaintext/cipher text pairs formed with the secret key and attempts to deduce the root key that used to produce this cipher text. In our approach, PSO is used as optimization technique to collect the optimal effective plaintexts from a plaintext search space according to the proposed fitness function then the set of collected plaintexts and the corresponding cipher texts used to extract the best eight sub-key differences from which most bits of the root key are discovered.

Rough sets and genetic algorithms: A hybrid approach to breast cancer classification

The use of computational intelligence systems such as rough sets, neural networks, fuzzy set, genetic algorithms, etc., for predictions and classification has been widely established. This paper presents a generic classification model based on a rough set approach and decision rules. To increase the efficiency of the classification process, boolean reasoning discretization algorithm is used to discretize the data sets. The approach is tested by a comparative study of three different classifiers (decision rules, naive bayes and k-nearest neighbor) over three distinct discretization techniques (equal bigning, entropy and boolean reasoning). The rough set reduction technique is applied to find all the reducts of the data which contains the minimal subset of attributes that are associated with a class label for prediction. In this paper we adopt the genetic algorithms approach to reach reducts. Finally, decision rules were used as a classifier to evaluate the performance of the predicted reducts and classes. To evaluate the performance of our approach, we present tests on breast cancer data set. The experimental results obtained, show that the overall classification accuracy offered by the employed rough set approach and decision rules is high compared with other classification techniques including Bayes and k-nearest neighbor.