Ayman Al-Dmour

Arabic writer identification based on hybrid spectral-statistical measures

Many techniques have been reported for handwriting-based writer identification. None of these techniques assume that the written text is in Arabic. In this paper we present a new technique for feature extraction based on hybrid spectral–statistical measures (SSMs) of texture. We show its effectiveness compared with multiple-channel (Gabor) filters and the grey-level co-occurrence matrix (GLCM), which are well-known techniques yielding a high performance in writer identification in Roman handwriting. Texture features were extracted for wide range of frequency and orientation because of the nature of the spread of Arabic handwriting compared with Roman handwriting, and the most discriminant features were selected with a model for feature selection using hybrid support vector machine–genetic algorithm techniques. Four classification techniques were used: linear discriminant classifier (LDC), support vector machine (SVM), weighted Euclidean distance (WED), and the K nearest neighbours (K_NN) classifier. Experiments were performed using Arabic handwriting samples from 20 different people and very promising results of 90.0% correct identification were achieved.

An evolving recommender-based framework for virtual learning communities

This article presents a recommender framework to provide personalised suggestions for learners taking introductory undergraduate courses. The framework utilises memory-based collaborative filtering algorithm combined with an imbedded web crawler to update learning material. The process of providing recommendation is divided into four steps: learner model extraction, neighbourhood formation, top-N recommendation presentation, and material update. The framework was implemented and has been successfully tested on real learners taking an introductory mathematics course. The learners of the course were divided into two groups. One of the groups, control group, was taught the material of the course using the traditional face-to-face approach. However, the students in the other group, experimental group, had the advantage to use the framework. The performance of the learners in both groups was tested and the results showed that the learners in the experimental group outperformed their counterparts in the control group.

Developing a virtual laboratory for a communication and computer networking course

Educational institutions face financial challenges in maintaining and upgrading traditional physical laboratories. Moreover, university students have challenges in understanding the theoretical issues involved in communication and networking courses. This paper presents a new computer-aided laboratory to help the students to enhance the practical view of communication and networking systems using computer modelling and simulation and to replace a physical laboratory. The new laboratory was build using a component-based approach that integrates modelling and simulation tools, namely MATLAB, SIMULINK, VPI and OPNET. These components were integrated through the development of graphical user interface. Through this interface, the student navigates through experiments of the laboratory. The new laboratory curricula and interface were tested on students and their feedback showed that the students were satisfied. The students gave comments and suggestions for future improvements.

An efficient compression technique using Lempel-Ziv algorithm based on dynamic source encoding scheme

This paper proposes an efficient text compression technique based on dynamic source encoding scheme using Lempel-Ziv algorithms. A statistical analysis is performed to count the occurrence of each character in the original source. Accordingly, each character is replaced by a weighted fixed-length code in lieu of arbitrary codes such as ASCII. This replacement generates an equivalent binary source with two symbols zero and one. Thereafter, the bitwise Lempel-Ziv algorithms can be applied to the nth-order extension binary source that contains 2 n symbols. As a result, the proposed technique achieves a high compression ratio in comparison with those techniques that use static and arbitrary source encoding schemes. In the decompression process, a header file includes all characters sorted in descending order based on their frequency of occurrences along with the extension order (n) is received. Thus, each character can be restored back based on the code shared between the compression and decompression processes.

An Evolutionary FIR Filter Design Method

In this introductory chapter, we present an evolutionary-based technique for designing one-dimensional and two-dimensional Finite Impulse Response (FIR) filters. Typically, the required filter has a given set of specifications to be met. The specifications may include cut-off frequency, band-stop region, band-pass region and ripple factors. The evolutionary method we are using is a modified version of the Genetic Algorithm (GA), which we call Flexible Genetic Algorithm (FGA). It is an optimization algorithm with high capabilities to span the space of filter parameters. FIR filters are highly required in different applications that process signals or images. A review of the state-of-the-art of filter optimization using evolutionary techniques is presented in this chapter. The aim of this work is to simply give a basic example of how filters can be designed using evolutionary techniques. As a matter of fact, medical applications require high linearity in the filter phase function to prevent undesired distortions in the detected signals. The proposed technique is based on minimizing a cost function that uses the weighted squared difference between the optimum filter specifications and the solutions generated by the evolutionary method. Comparisons between FGA-designed filter and standard method designed filters are implemented. Testing of filters is done using different noisy artificial ECG signals and selected images. We used Hermite functions to build the artificial ECG signals.

An Efficient Bit-Level Lossless Grayscale Image Compression Based on Adaptive Source Mapping

Image compression is an essential technique for saving time and storage space for the gigantic amount of data generated by images. This paper introduces an adaptive source-mapping scheme that greatly improves bit-level lossless grayscale image compression. In the proposed mapping scheme, the frequency of occurrence of each symbol in the original image is computed. According to their corresponding frequencies, these symbols are sorted in descending order. Based on this order, each symbol is replaced by an 8-bit weighted fixed-length code. This replacement will generate an equivalent binary source with an increased length of successive identical symbols (0s or 1s). Different experiments using Lempel-Ziv lossless image compression algorithms have been conducted on the generated binary source. Results show that the newly proposed mapping scheme achieves some dramatic improvements in regards to compression ratios.