|
Home>Products>Journals
and Reviews>I.Re.Co.S.>Latest
issue
International Review
on Computers and Software - Papers
 An
Unsupervised Classification Using a Novel ICM Method with Constraints for Land
Cover Mapping from Remote Sensing Imagery
by S. Idbraim, D.
Ducrot, D. Mammass, D. Aboutajdine
Abstract -
The use of contextual
information for modeling the prior probability function has found
applications in remotely sensed data classification. The Markov Random
Fields (MRF) provide a theoretically robust mathematical tractable way
of modeling contextual behavior. In this paper we present a MRF method
for classification of satellite images, this method is based on
minimization of the energy by the ICM method (Iterated Conditional
Mode). The originality of the proposed method is the introduction of:
firstly, a parameter of temperature such as in the Simulated Annealing
algorithm (SA), and secondly, a new contextual constraint which is the
segmentation contours. To choice the reliable map of contours, we have
developed automatic method for the selection of threshold. This
algorithm refines the classification by re-estimating the statistics of
classes according to the previous iteration and by giving more
importance to the contextual information through the parameter of
temperature. A Landsat ETM+ image is used to illustrate the
effectiveness of the proposed method for an application of cartography.
The results show a significant increase in the accuracy of land cover
maps comparing to other classification methods.
Copyright © 2009 Praise
Worthy Prize S.r.l. - All rights reserved
Keywords:
Classification, Iterated Conditional Mode, Land cover, Markov
Random Fields, Satellite images, Segmentation.
Planar
Nuclear Images Restoration Based on Fourier and Wavelet Transforms
by Nawres Khlifa,
Nejla Gribaa, Kamel Hamrouni
Abstract - Nuclear
images (called also scintigraphic images) are often disturbed by blur
and noise. This is due especially to the equipment quality and the
disintegration of the radioactive substance injected to the body. The
region of interest perception is then a very difficult task. In this
context, several soft frameworks are developed in order to overcome
these limitations. Some works focused on denoising nuclear images;
others consider the problem as a general restoration one. The
restoration is an ill-posed problem, so the direct deconvolution of the
distortion model is often numerically unstable. In this paper, we
propose a new framework that takes benefits jointly of the Fourier and
the Wavelet domains advantages. It consists in operating firstly by a
regularized deconvolution using the Fourier basis, and regularizing
secondly the deconvolved image by shrinkage in the Wavelet packet basis.
The method provides good results in terms of edge preservation, contrast
and uniformity in the images.
Copyright © 2009 Praise Worthy Prize S.r.l. -
All rights reserved
Keywords:
Nuclear medecine, Image restoration, Regularisation, Wavelet packets
transform, Fourier transform.
Interactive
Three-Dimensional Segmentation of MR Images by Hierarchical Watershed
by Jlassi
Hajer, Hamrouni Kamel
Abstract - This
paper describes a fast method of interactive 3D segmentation using the
hierarchical techniques of the Watershed [1]. This method uses the
hierarchical queues watershed algorithm as well as a merging process to
limit over segmentation. The waterfall hierarchical 3D algorithm is fast
when constructing the saddle point graph. However, it can be noticed
that too many areas are merged from a hierarchical level to the next
one. That is why a hybrid algorithm, combining waterfall and
hierarchical merging approaches is proposed. It allows, on a criterion
of number of wished regions, to obtain a fast and effective segmentation
on 3D image for which the user has an idea of the volume to be
extracted. Our method has been applied on a data base of 3D brain MRIs.
Results are illustrated by segmenting the ventricles, corpus callosum,
cerebellum, hippocampus, pons, medulla and midbrain on web-brain data
base and the approach is validated on 20 Normal Subjects, T1-Weighted
Scans. This validation showed the robustness of our method in detection
of the volumes.
Copyright © 2009 Praise Worthy Prize S.r.l. -
All rights reserved
Keywords:
3D Segmentation, 3D Watersheds, Waterfall, Merging.
A
Method for Automatic Score Box Detection and Text Recognition in Soccer
Video
by
Youness Tabii, Rachid Oulad Haj Thami
Abstract - This Text
in video is normally generated in order to supplement or summaries’ the
visual content and thus is an important carrier of information that is
highly relevant to the content of the video. As such, it is a potential
ready-to-use source of semantic information. In this paper we present a
new algorithm for detection of score box superimposed on the soccer
video. This method is based on motion vector. After score box detection,
the next step is the recognition of text that exists in this score box
using morphological processing and the optical character recognition
(OCR).
Copyright © 2009 Praise Worthy Prize S.r.l. -
All rights reserved
Keywords:
Text Extraction, Optical Character Recognition (OCR), Soccer Video,
Binarization, Motion Vector.
by Fateh
Boutekkouk, Sebastien Bilavarn, Michel Auguin, Mohamed Benmohammed
Abstract - SystemC
is considered as a de facto open source standard for Systems On Chip
(SOC) system level design. It supports a diversity of untimed and timed
models at different levels of abstraction, ranging from high-level
functional models to detailed clock cycle accurate RTL (Register
Transfer Level) models and iterative refinement of high level models
into lower levels of abstraction. Despite its wide spread use, SystemC
which is basically a C++ library lacks an adequate formal semantics. On
the other side, Rewriting Logic (RL) is an algebraic formalism which
allows the dynamic behavior of systems to be axiomatised using rewriting
rules. RL has been used to develop an object-oriented specification and
verification language Maude for distributed and concurrent systems. In
this paper, we demonstrate the expressive power and versatility of Maude
by applying it to the definition of a formal semantics for SystemC
scheduler.
Copyright © 2009 Praise Worthy Prize S.r.l. -
All rights reserved
Keywords:
SystemC, Rewriting Logic, Maude, Formal Verification.
Phoneme
Classification Using Different Approaches Based on Fuzzy Measurements and Neural
Networks
by D. Ben Ayed, N.
Arous, N. Ellouze
Abstract - Speech
recognition is a difficult problem due to the inability of current
systems to cope with connected speech. Neural networks and fuzzy logic
are able to learn some aspects of this task. An unsupervised learning
neural network scheme like self-organizing map (SOM) can be used to both
classify and order speech sounds. SOM provides a means of reducing
inherent dimensionality of speech data. Fuzzy logic has been applied
very successfully in many areas where conventional model based
approaches are difficult or not cost effective to implement. Fuzzy
similarity measurement was used for approximate reasoning as it attempts
to model the human reasoning process. In this paper, we propose SOM
variants and fuzzy measurements in order to improve phoneme
classification accuracy. The proposed SOM with locally neighbourhood
radius (AdSOM) and fuzzy system based on Wang similarity (MSFLwen2) show
good robustness and high phoneme classification rates in the context of
phonetically near phonemes.
Copyright © 2009 Praise Worthy Prize S.r.l. -
All rights reserved
Keywords:
Neural network, Self-organizing map, SOM learning variant, Fuzzy
logic, Fuzzy measurement, Fuzzy similarity, Phoneme classification.
Classifiers
Selection and Features Extraction / Selection for Arabic Handwritten
Word Recognition
by
Azizi Nabiha, Sellami Mokhtar
Abstract - The
performance of multiple classifier systems (MCS) depends essentially on
the complementarities (diversity) of constitutes single classifiers. The
complementarity can be obtained by diversifying the feature
representation, classifier structure, as well as other factors. The
author desire is to state the basics along with a survey for solving
Arabic handwritten recognition problems. This paper focuses on the
parameters choice for the two paradigms of features and multi
classifiers systems; this is done proposing two approaches. The first
rests on cooperation and selection of feature set in MCS, while the
second targets the selection of best classifier sub set. For the latter,
three levels of problems are debated: The nature of classifiers used,
the manner upon which a set of classifiers is selected using a diversity
measures between classifiers and the adopted combination method. The
experimental results presented are encouraging and open other
perspectives in the domain of the features and classifier selection. The
multiple classifiers systems (MCS).
Copyright © 2009 Praise Worthy Prize S.r.l. -
All rights reserved
Keywords:
Arabic Handwritten Word, Set of Classifier, Diversity Measures, Fusion
Classifier, Feature Selection/Extraction, Voting, Weighted Voting, BKS (Behaviour
Knowledge Space).
WordNet-Based
Text Clustering Methods: Evaluation and Comparative Study
by A. Amine, Z.
Elberrichi, M. Simonet
Abstract - Text
clustering is one of the most important applications of Text Mining.
Several studies have been conducted in this field; however, no algorithm
can claim a universal effectiveness. Indeed, a clustering method can be
adapted to a certain type of situations, and be useless when faced with
certain other categories of data. This explains the proliferation of
different and varied approaches. The success of a text clustering method
depends on a complex set of parameters that are the choice of corpus and
their variants, coding documents, the choice of similarity measures
between these documents and the choice of algorithms and their
implementation. It is therefore extremely difficult to draw definitive
conclusions on an approach. In this paper we present an evaluation and a
comparison of three text clustering methods based on WordNet for the
representation of texts. Several experiments using different similarity
measures and the Reuters-21578 benchmark corpus are conducted.
Copyright © 2009 Praise Worthy Prize S.r.l. -
All rights reserved
Keywords:
Text clustering, Similarity, WordNet, Reuter-21578.
On
Global Clustering Algorithm: Layer-Oriented Approach for First/Last Node
Dying Applications in Wireless Sensor Networks
by S. M. Mazinani,
M. H. Yaghmaee, F. Tashtarian, M. T. Honary, J. Chitizadeh
Abstract -
Clustering methods are potentially the framework for power-conserving
wireless sensor networks. These methods are among the most effective
approaches aiming for prolonging the lifetime. Power-limited nature of
the employed devices in wireless networks made the conventional
clustering approaches to mainly focus on selecting cluster heads -which
are the central points within a cluster- among the nodes with more
remaining energy. Periodic rotation of this task -being a cluster head-
between nodes leads to distribute the energy consumption. In singlehop
wireless sensor network, Hot-Spot problem arises when the further
cluster heads to the base station excessively utilize their energy due
to the distant communications. The main purpose of this paper is to
provide a layering-based clustering algorithm that aims at directly
optimizing the nodes’ longevity by considering the application in which
they operate and the relationship between the width of each layer and
its overall lifetime. Two main types of applications are dealt with;
First node dying and Last node dying applications. Analytical methods
are employed to study the impacts of intra-cluster and singlehop
communication traffic loads on the average layer’s lifetime that is
formulated as an optimization problem.
Copyright © 2009 Praise Worthy Prize S.r.l. -
All rights reserved
Keywords:
Wireless Sensor Networks, Clustering, Layering, Lifetime, Single Hop.
A
New Middleware for Multipurpose Applications in Wireless Sensor Networks
by S. M. Mazinani,
M. H. Yaghmaee, M. T. Honary1,2, F. Tashtarian4, J. Chitizadeh
Abstract - Dramatic
development of the wireless sensor networks (WSNs) in several
applications is seeking for multi-application implementation of such
networks. In this way, the middlewares contain a runtime environment
that supports and coordinates multiple applications. In this paper, a
middleware is proposed to support multi-application network that
simultaneously satisfies the quality of service (QoS) requirements of
users, that is the response time and data accuracy, and that of network
that is optimizing the network resources (i.e. energy) utilization.
Organizing nodes into clusters and preparing them to efficiently report
data of any detected event, the middleware features an inheritance
pattern among nodes that trains them to support multiple events. More
over, employing the Data Integration among nodes not only reduces the
energy consumptions for data reporting but also increases the data
accuracy of every detected event.
Copyright © 2009 Praise Worthy Prize S.r.l. -
All rights reserved
Keywords:
Middleware, Inheritance, Wireless Sensor Network, Clustering,
Quality of Service, Data Accuracy.
Unit
Testing of Mobile Agents Using the Nets within Nets Paradigm
by Yacine Kissoum,
Zaďdi Sahnoun, Kamel Barkaoui
Abstract - Among all
the different architectures being researched in the field of multi-agent
systems, the mobile agent has shown to be one of the most challenging
and most critical systems. With more applications being developed, there
is a need to ensure large and complicated mobile agent systems are
functioning correctly, with minimum or no errors. Moreover, the
model-based testing technique has gained attention with the
popularization of models in software design and development. Since the
paradigm on nets within nets is well suited to express the dynamics of
open mobile agents, it is retained as an abstract model from which
abstract test cases are generated. Those test cases are then concretized
and addressed to the system under test. The responses of the system
under test are, finally, compared to the expected results derived from
the abstract test model. As a case study, we modeled a packet world
example on which different colored packets are scattered. Agents that
live in this virtual world have to collect those packets and bring them
to their right destination.
Copyright © 2009 Praise Worthy Prize S.r.l. -
All rights reserved
Keywords:
Mobile Agents, Nets within Nets, Reference Nets, Renew, Testing.
Modeling Method of Robust Control Laws for Manufacturing System to
Temporal and non Temporal Constraints through Petri Nets
by Achraf Jabeur
Telmoudi, Lotfi Nabli, Radhi M’hiri
Abstract - This
article aims to establish a formal framework for modeling the robust
control laws of manufacturing systems to temporal and non temporal
constraints. The goal is to conclude the type from the production system
robustness. A methodology of robust control construction generating the
margins of passive and active robustness is elaborated. The redundancy
of the robustness of the elementary parameters between passive and
active is used. Indeed, local models are developed; each of them
modeling resource robustness. Then, a synchronized assembly of these
models enabled us to constitute our final model. The Intervals
Constrained Petri Nets (ICPN) tool is used. To this goal, a whole of
definitions, lemmas and theorems are committed and affirmed per
applicable examples. An application to a tobacco manufacturing is
presented.
Copyright © 2009 Praise Worthy Prize S.r.l. -
All rights reserved
Keywords:
Manufacturing systems control, Temporal constraints, Non temporal
constraints, Robustness, Modeling, Petri nets.
Web
Environment for Psychometric Diagnostics and Psychotherapy
by Velibor Ilić
Abstract - In this
paper is described web environment based on content management system
that can integrate several tools and subsystems specialized for
psychometric diagnostic, therapy and communication. Environment provides
collecting information about users, evaluation cognitive capabilities,
interpretation and visualization of test results, database analyzes, and
generation of user’s profile and their cognitive maps. Adaptive
multimedial diagnostics tests integrated in this environment can be used
for evaluation of user’s cognitive functions. Data mining allows
processing the large amount collected information (diagnostic test
results and information about users) and it reports about potentially
interesting data. Environment provides online platform where patients
and field experts (psychologists/therapists) can organize in virtual
communities. Advantages of this environment can be seen from perspective
of psychologists and persons who are going to be tested.
Copyright © 2009 Praise Worthy Prize S.r.l. -
All rights reserved
Keywords:
Framework for Integrated Testing (FIT), Content Management Systems
(CMS), Computer Supported Cooperative Work (CSCW), Collaborative Working
Environment (CWE), Workgroup Support Systems (WSS), Virtual Communities,
Psychometric Diagnostic.
User
Localization in Cellular Mobile Environment
by Slama Zohra
Abstract -
In the mobile networks users that dispose of mobile units can move from
one location to another and then living their home network. The
objective in mobile computing is to have the possibility of working and
using the network resources during the moves of the users without
geographic constraints. However, a large number of problems, materials
or software, must be solved. We consider in this paper the users
location management problem that is fundamental in the query processing.
The mobile users can communicate with other static or mobiles users
regardless of its current location and mobility pattern. To reach this
objective, it is important to determine the current location of a
mobile. A location management scheme is then necessary. In this paper,
we use an approach based on the prediction and we propose a parallel
strategy to localize mobile units in a network.
Copyright © 2009 Praise Worthy Prize S.r.l. -
All rights reserved
Keywords:
Mobile Network, Cell, Localization, Database, Hierarchical
Network, Optimization, Queries, Parallelism.
G.I.S.
Contribution for Study the Risks Zones in Urban Area. The “Oran-Risk”
by D. D. Rahal, F.
Z. Baba Hamed, F. Rahal
Abstract - The civil
engineering structures study consists in dealing with any abnormal
phenomenon (seism, landslide, flood,etc.) could to appear in the urban
area to control the urban extension in the exposed zones by reconciling
the prevention requirements and the social and economic development
needs. The geological characteristics, the geographical situation and
the socio-economic importance of Oran City, form a great and complex
system where the risk catastrophes is great. This study proposes to
determine these risks zones by using the Geographical Information System
(G.I.S.). This system is developed to produce thematical maps
corresponding to the different hazards types in the urban area. The
latter recombined with the structural vulnerabilities (buildings and
structures vulnerability) will make it possible to produce maps of risk.
A G.I.S named Oran-Risk was developped. This tool can prove to be a
basic element in the evolution of the integrated territory management of
the dangers compared and aid the decision to make the construction rules
and the site urbanisable choice.
Copyright © 2009 Praise Worthy Prize S.r.l. -
All rights reserved
Keywords:
Methodology, Risks zones, G.I S., Risks, Vulnerability, Making
Decision Aid.
| 
