Annual Indexing

Aspects of Artificial Neural Networks as a Modelling Tool for Industrial Processes

Mohamed Khadir

Department of Computer Science, University Badji Mokhtar of Annaba, Algeria

Abstract: In order to investigate the behavior of industrial processes for design, fault prevention, prediction, control, etc., a model of the process is necessary. Due to inherent nonlinearities proper to industrial processes, and/or nonlinearities due to the characteristics of the valves and pumps forming the entire industrial plant, nonlinear models are desired. Complete mathematical models of such plants proves to be time and efforts consuming, when not totally unrealizable. The fact that Artificial Neural Networks (ANNs) have been proven, by Cybenko, able to represent any nonlinear function, as well as their easy implementation, led to their widespread usage in the modeling community; often not at best and ending in controversial results. This paper proposes a methodology for designing and validating ANN models for modeling industrial plants, taking into consideration typical industrial constraints such as restricted data sets. The approach is applied to an industrial milk pasteurization plant.

Keywords: Artificial neural networks, multi-layer perceptron, nonlinear models, pasteurization plant.

Received July 10, 2004; accepted September 17, 2004

Generalized Hough Transform for Arabic Printed Optical Character Recognition

Sofien Touj¹, Najoua Ben Amara², and Hamid Amiri¹

¹National Engineers School of Tunis, Tunisia

² National Engineers School of Monastir, Tunisia

Abstract: The Hough Transform (HT) is a technique commonly used in image processing. It is known for its capacity to detect objects in a given image. In the present paper, we propose to explore the properties of the HT and the use of the Generalized HT (GHT) in Arabic Optical Character Recognition (AOCR). Hence, we first present a GHT based approach for the recognition of Arabic printed characters in their different shapes depending on their position in the word. Accordingly character models are stored in a structure called dictionary which is used further for text recognition. In fact, we have proposed two segmentation-by-recognition techniques for cursive printed writing recognition. The first one uses a technique by a dynamic sliding window. The second one is based on the identification and the localisation of the characters within a word or a part of a word called also sub word. Some outcomes of this study are also assessed in this paper.

Keywords: Generalized Hough transform, Arabic printed optical character recognition, printed cursive writing, segmentation by recognition techniques.

Received July 6, 2004; accepted November 1, 2004

Arabic Handwritten Words Recognition Based on a Planar Hidden Markov Model

Sameh Touj¹, Najoua Ben Amara², and Hamid Amiri¹

¹National Engineers School of Tunis, Tunisia

² National Engineers School of Monastir, Tunisia

Abstract: Off-line recognition of handwritten words is a difficult task due to the high variability and uncertainty of human writing. The majority of the recent systems are constrained by the size of the lexicon to deal with and the number of writers. In this paper, we propose an approach for multi-writers Arabic handwritten words recognition. The developed method uses multiple sources of information at the description and the classification levels. A hybrid planar Markovien modelling permitting to follow the horizontal and vertical variations of the writing has been adopted. This modelling is based on different levels of segmentation: horizontal, natural and vertical. The process of segmentation conducts to the decomposition of the writing in a limited set of elementary entities, with simplified morphologies specific to every horizontal band. The choice of different type of primitives is then imposed in order to assure an efficient description. Different architectures of modelling proved also to be indispensable. The classification is finally achieved using a Planar Hidden Markov Model.

Keywords: Off-line Arabic handwriting recognition, planar hidden Markov models, segmentation, multiple sources of information.

Received July 2, 2004; accepted September 17, 2004

Design and Evaluation of an Input Buffered Packet Switch

Azeddine Bilami¹, Mustapha Lalam², Mehammed Daoui², and Mohamed Benmohammed³

¹ Department of Computing Science, University of Batna, Algeria

² Department of Computer Science, University of Tizi Ouzou, Algeria

³ Department of Computer Science, University of Constantine, Algeria

Abstract: Many architectures of internet routers, ATM and ethernet switches have been proposed and analysed in literature. Theoretically reliable and valid solutions have been developed to achieve high performances but a lot of them are not feasible in practice for commercial and technological reasons. Few papers develop the implementation and simulation aspects. The objective of this paper is the design of a packet switch with a minimum cost and hardware complexity. We propose an input-queuing architecture using a multistage interconnection network and a simple cell selection policy implemented by hardware. The switch is described and simulated using a VHDL language. Performances in terms of throughput and cell loss are evaluated.

Keywords: Routing, switch, multistage interconnection network, Benes network, self routing, VHDL.

Received June 23, 2004; accepted October 30, 2004

Integrating Software Traceability for Change Impact Analysis

Suhaimi Ibrahim¹, Norbik Bashah Idris¹, Malcolm Munro², and Aziz Deraman³

¹ Centre for Advanced Software Engineering, University of Technology Malaysia, Malaysia

² Department of Computer Science, University of Durham, United Kingdom

³ Faculty of Technology and Information System, University Kebangsaan Malaysia, Malaysia

Abstract: Software maintenance is recognized as the most costly activity in software engineering with typical estimates of more than half of the software development cost. The main problem to a maintainer is that seemingly small changes can ripple throughout the system to cause substantial impact elsewhere. Software traceability and its subsequent impact analysis help relate the consequences or ripple-effects of a proposed change across different levels of software models. In this paper, we present a software traceability approach to support change impact analysis of object oriented software. The significant contribution in our traceability approach can be observed in its ability to integrate the high level with the low level software models that involve the requirements, test cases, design and code. Our approach allows a direct link between a component at one level to other components at any levels. It supports the top down and bottom up traceability in response to tracing for the ripple-effects. We developed a software prototype called Catia to support C++ software, applied it to a case study of an embedded system and discuss the results.

Keywords: Requirement traceability, impact analysis, concept location, change request.

Received June 17, 2004; accepted October 30, 2004

Genetic-Neural Approach versus Classical Approach for Arabic Character Recognition Using Freeman Chain Features Extraction

Raed Abu Zitar

College of Information Technology, Philadelphia University, Jordan

Abstract: This article presents a hybrid technique for the recognition of typed Arabic characters. Due to its curved and continuous nature, Arabic text has to go through words segmentation, character segmentation, feature extraction, and finally character recognition. In this work, Freeman Chain (FC) technique [20, 21] is used to generate a chain for every segmented character. This chain represents the extracted features. Moreover, two approaches are presented for the classification process. In the first approach, we use a classical sequential weighing algorithm that finds the closest available “Standard Character Template” to the extracted chain. In the second approach, we use Learning Vector Quantization (LVQ) (specifically LVQ3) technique for classifying the same chain. To improve the performance of that LVQ, the Genetic Algorithm (GA) [11, 23] is invoked for some additional training. We call our neural network with the GA “GALVQ3”. For further robustness testing of both approaches, we add some artificial noise to the extracted chains and repeat simulations. In general, LVQ techniques provide higher classification rate even for cases where noise and partial observations exist. As a result, the GALVQ3 classifier is compact, online, robust, and feasible from hardware point of view.

Keywords: Arabic character recognition, neural networks, Freeman chain, feature extraction, LVQ.

Received June 6, 2004; accepted October 4, 2004

Minimum Cost Path for a Shared Nothing Architecture

Maytham Safar

Computer Engineering Department, Kuwait University, Kuwait

Abstract: Computing the minimum cost path is a key requirement in Intelligent Transportation Systems (ITS) and in some Geographical Information Systems (GIS) applications. The major characteristics of these systems are the facts that the underlying transportation graph is large in size and the computation is under time constraint. Due to the insufficiency of the classic algorithms under these settings, recent studies have focused on speeding the computation by employing alternative techniques such as heuristics, precomputation and parallelization. In this study, we investigate solutions assuming a shared nothing architecture (i. e., Teradata multimedia database system) as a way of speeding up the computation further. We build our algorithms on a recently developed graph model, Hierarchical mulTigraph (HiTi), and describe both concurrent and parallel versions of the algorithms. The concurrent algorithm allows simultaneous exploration of the search space by utilizing dynamically created agents across multiple disk nodes, which is efficiently supported by the Teradata multimedia database system architecture. The parallel algorithm breaks the problem into a set of smaller subproblems by exploiting a set of intermediate nodes that the shortest path passes through. We also investigate the impact of replicating subgraphs in the performance of our algorithms. We evaluated our algorithms via a simulation study and demonstrated that our concurrent and parallel algorithms show almost a linear speedup as the number of disk/CPU nodes is increased. Concurrent algorithm exhibits better sizeup, and scaleup results than the parallel algorithm.

Keywords: Shortest path, GIS system, intelligent transportation systems, shared nothing architecture, teradata multimedia database.

Received May 24, 2004; accepted July 31, 2004

Implementing Temporal Radial Basis Function for Reactive Navigation of Mobile Robot

Mesbahi Larbi ¹, Hendel Fatiha ², Berrached Nasr-Eddine², and Abdelkader Benyettou ¹

¹ Laboratoire SIMPA, University of Sciences and Technology of Oran, Algeria

² Laboratoire LARESI, University of Sciences and Technology of Oran, Algeria

Abstract: We present in this article, the realisation of reactive navigation module based on neural networks Like Temporal Radial Basis Functions (TRBF), with respect of security constraints and inherent robustness while using an Orthogonal Least Square algorithm (OLS). Applied to a structured type like interior of building, the mobile robot must assure its task of navigation mildly all while avoiding obstacles without wandering, with the possibility to take into account the taken decisions in its past lasting trajectory.

Keywords: Reactive navigation, TRBF, OLS, mobile robot.

Received May 18, 2004; accepted August 15, 2004

Comparison of the Hardware Implementation of Stream Ciphers

Michalis Galanis, Paris Kitsos, Giorgos Kostopoulos, Nicolas Sklavos, and Costas Goutis

Electrical and Computer Engineering Department, University of Patras, Greece

Abstract: In this paper, the hardware implementations of five representative stream ciphers are compared in terms of performance and consumed area in an FPGA device. The ciphers used for the comparison are the A5/1, W7, E0, RC4 and Helix. The first three ones have been used for the security part of well-known standards, especially wireless communication protocols. The Helix cipher is a recently introduced fast, word oriented, stream cipher. W7 algorithm has been recently proposed as a more trustworthy solution for GSM, due to the security problems concerning A5/1. The designs were implemented using VHDL language. For the hardware implementation of the designs, an FPGA device was used. The implementation results illustrate the hardware performance of each stream cipher in terms of throughput-to-area ratio. This ratio equals to: 5.88 for the A5/1, 1.26 for the W7, 0.21 for the E0, 2.45 for the Helix and 0.86 for the RC4.

Keywords: Cryptography, security, stream ciphers, hardware architecture, FPGA implementation.

Received April 27, 2004; accepted July 28, 2004

ARABASE: A Relational Database for Arabic OCR Systems

Najoua Ben Amara ¹, Omar Mazhoud ¹, Noura Bouzrara ², and Noureddine Ellouze ²

¹ National School of Engineer of Monastir, Tunisia

² National School of Engineer of Tunis, Tunisia

Abstract: In this paper we present a database for the research of Arabic off-line and on-line handwriting optical recognition as well as for machine printed text optical recognition. Digital images of documents, text phrases, words/sub-words, isolated characters, digits, signatures, soon are and included in ARABASE. Data corresponds to a variety of lexes (cities names, literal amounts, isolated characters, digits, free texts, etc.). The database organization offers interesting commodities to be explored via an Arabic writing recognition system. A useful tool enables the user, via a graphical interface to experiment different classical tasks of image processing.

Keywords: Databases, Arabic writing recognition, on-line and off-line handwriting, printed documents, multi-fonts, multi-writers.

Received July 2, 2004; accepted September 17, 2004