SICISI Unit

Tunis, Tunisia

SICISI Unit

Tunis, Tunisia
SEARCH FILTERS
Time filter
Source Type

Ben Chaabane S.,SICISI Unit | Sayadi M.,SICISI Unit | Sayadi M.,University of Picardie Jules Verne | Fnaiech F.,SICISI Unit | And 2 more authors.
Eurasip Journal on Advances in Signal Processing | Year: 2010

A novel method of colour image segmentation based on fuzzy homogeneity and data fusion techniques is presented. The general idea of mass function estimation in the Dempster-Shafer evidence theory of the histogram is extended to the homogeneity domain. The fuzzy homogeneity vector is used to determine the fuzzy region in each primitive colour, whereas, the evidence theory is employed to merge different data sources in order to increase the quality of the information and to obtain an optimal segmented image. Segmentation results from the proposed method are validated and the classification accuracy for the test data available is evaluated, and then a comparative study versus existing techniques is presented. The experimental results demonstrate the superiority of introducing the fuzzy homogeneity method in evidence theory for image segmentation. © 2010 Salim Ben Chaabane et al.


Balti A.,SICISI Unit | Sayadi M.,SICISI Unit
Journal of Electrical Engineering | Year: 2013

This paper is concerned with novel features for fingerprint classification based on the Euclidian distance between the center point and their nearest neighbor bifurcation minutia's. The main advantage of the new method is the dimension reduction of the features vectors used to characterize fingerprint, compared with the classic characterization method based on the relative position of bifurcation minutia points. In addition, this new method avoids the problem of geometric rotation and translation over the acquisition phase. Whatever, the degree of fingerprint rotation, the extraction features used to characterize the fingerprint remains the same. The characterization efficiency of the proposed method is compared to the method based on the spatial coordinate position of fingerprint minutia's. The comparison is based on a characterization criterion, usually used to evaluate the class quantification and the features discriminating ability. After that, the classification accuracy of the proposed approach is evaluated with Back Propagation Neural Network (BPNN). Extensive experiments prove that the fingerprint classification based on a novel features and BPNN classifier give better results in fingerprint classification than several other features and methods.


Balti A.,SICISI Unit | Sayadi M.,SICISI Unit | Fnaiech F.,SICISI Unit | Fnaiech F.,University of Picardie Jules Verne
2013 10th International Multi-Conference on Systems, Signals and Devices, SSD 2013 | Year: 2013

This paper is concerned with novel features for fingerprint classification based on the Euclidian distance between the center point and their nearest neighbor bifurcation minutia's. The main advantage of the new method is the dimension reduction of the features vectors used to characterize fingerprint, compared with the classic characterization method based on the relative position of bifurcation minutia points. In addition, this new method avoids the problem of geometric rotation and translation over the acquisition phase. The characterization efficiency of the proposed method is compared with the method based on the spatial coordinate position of fingerprint minutia's. The comparison is based on a characterization criterion, usually used to evaluate the class quantification and the features discriminating ability. After that, the classification accuracy of the proposed approach is evaluated with Back Propagation Neural Network (BPNN). Extensive experiments prove that the Fingerprint classification based on a novel features and BPNN classifier give better results in fingerprint classification than several other features and methods. © 2013 IEEE.


Tlig L.,SICISI Unit | Sayadi M.,SICISI Unit | Fnaeich F.,SICISI Unit
2010 2nd International Conference on Image Processing Theory, Tools and Applications, IPTA 2010 | Year: 2010

In this paper we present a novel segmentation approach that performs fuzzy clustering and feature extraction. The proposed method consists in forming a new descriptor combining a set of texture sub-features derived from the Grating Cell Operator (GCO) responses of an optimized Gabor filter bank, and Local Binary Pattern (LBP) outputs. The new feature vector offers two advantages. First, it only considers the optimized filters. Second, it aims to characterize both micro and macro textures. In addition, an extended version of a type 2 fuzzy cmeans clustering algorithm is proposed. The extension is based on the integration of spatial information in the membership function (MF). The performance of this method is demonstrated by several experiments on natural textures. ©2010 IEEE.


Marzouki A.,SICISI Unit | Hamouda M.,SICISI Unit | Fnaiech F.,SICISI Unit
IEEE International Symposium on Industrial Electronics | Year: 2010

The main objective of this paper is to implement a nonlinear control for a PWM voltage source converter with a reduced number of sensors. For this purpose an input-output feedback linearization control strategy is firstly implemented in order to control both the line current and the DC output voltage. Next, an estimator of the AC mains voltage is implemented so as to reduce the number of the requested sensors, to avoid the undesired noise, and to minimize the converter's cost. The efficiency of the proposed control law and estimation method is next validated through simulation results. Consequently, load independent unity input power factor operation and a perfect tracking of the DC bus voltage waveform are achieved even without using mains voltage sensors. © 2010 IEEE.


Zaafouri A.,SICISI Unit | Sayadi M.,SICISI Unit | Fnaiech F.,SICISI Unit
2nd International Conference on Communications Computing and Control Applications, CCCA 2012 | Year: 2012

This paper presents a method of isolated Arabic character recognition using local energy and structural features. The method requires skeletonization in order to facilitate feature extraction process. The thinned character image is convolved with log Gabor filters bank for local energy feature extraction. Also, structural features such as: dots, endpoints and loops, are extracted from the skeleton to make easy the recognition stage. The characters are classified and recognized using multiplayer perceptron neural network MLP. Simulation results prove that the proposed set of features gives satisfactory recognition rate. Also the recognition system using local energy model demonstrates its rotation, scale and translation invariant. © 2012 IEEE.


Trabelsi O.,SICISI Unit | Tlig L.,SICISI Unit | Sayadi M.,SICISI Unit | Fnaiech F.,SICISI Unit
2013 International Conference on Electrical Engineering and Software Applications, ICEESA 2013 | Year: 2013

Tracking of the skin disease is a necessary step of diagnostic as well the measure of the wound's surface is very useful in healing's document. To overcome the difficulties of the skin illness's estimation, encountered with the currently used measurement techniques, we propose a novel approach aiming to reduce the time-consuming and the error rate. The proposed method is based on two steps; the first step is a preprocessing one which consists in image segmentation to detect the edge of the infected skin region. In the second one, another proposed method is applied to measure the wound 'size' and control the illness evolution. In this work, a comparative study was realized to select the most suitable segmentation technique referred to a proposed criterion based on 'edge accuracy' EAC. The new criterion was compared with the 'surface accuracy' based on ROC space. The experiments show the performance of the proposed criterion and the efficacy of the measurement technique. © 2013 IEEE.


Tlig L.,SICISI Unit | Sayadi M.,SICISI Unit | Fnaiech F.,SICISI Unit
Signal Processing: Image Communication | Year: 2012

Gabor filtering is a widely applied approach for texture analysis. This technique shows a strong dependence on certain number of parameters. Unfortunately, each variation of values of any parameter may affect the texture characterization performance. Moreover, Gabor filters are unable to extract micro-texture features which also have a negative effect on the clustering task. This paper, deals with a new descriptor which avoids the drawbacks mentioned above. The novel texture descriptor combines grating cell operator outputs derived from a designed Gabor filters bank, and local binary pattern features. For the clustering task, an extended version of fuzzy type 2 clustering algorithm is also proposed. The effectiveness of the proposed segmentation approach on a variety of synthetic and textured images is highlighted. Several experimental results on a set of textured images show the superiority of the proposed approach in terms of segmentation accuracy with respect to quantitative and qualitative comparisons. © 2012 Elsevier B.V. All rights reserved.


Zaafouri A.,SICISI Unit | Sayadi M.,SICISI Unit | Fnaiech F.,SICISI Unit
International Multi-Conference on Systems, Signals and Devices, SSD'11 - Summary Proceedings | Year: 2011

In this paper, we propose a developed unsharp masking process for contrast image enhancement. The main idea here is to enhance the dark and bright area in the same way which matches the response of human visual system well. Then in order to reduce the noise effect, a mean weighted high pass filter is used for edge extraction. The proposed method gives satisfactory results for wide range of low contrast images compared with others known approaches. © 2011 IEEE.


Balti A.,SICISI Unit | Sayadi M.,SICISI Unit | Fnaiech F.,University of Picardie Jules Verne
Control Engineering and Applied Informatics | Year: 2013

This paper is concerned with novel features for fingerprint classification based on the Euclidian distance between the center point and their nearest neighbor bifurcation minutiae's. The main advantage of the new method is the dimension reduction of the features vectors used to characterize fingerprint, compared with the classic characterization method based on the relative position of bifurcation minutiae points. In addition, this new method avoids the problem of geometric rotation and translation over the acquisition phase. Whatever, the degree of fingerprint rotation, the extraction features used to characterize fingerprint remains the same. The characterization efficiency of the proposed method is compared to the method based on the spatial coordinate of fingerprint minutiae's. The comparison is based on a characterization criterion, usually used to evaluate the class quantification and the features discriminating ability. After that, the classification accuracy of the proposed approach is evaluated with Back Propagation Neural Network (BPNN). Extensive experiments prove that the fingerprint classification based on a novel features and BPNN classifier gives better results in fingerprint classification than several other features and methods. Finally the results of the proposed method are evaluated on the FVC 2002 database.

Loading SICISI Unit collaborators
Loading SICISI Unit collaborators