Noureddine Ellouze

Tunis El Manar University

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Trends

Author Order

Document Type

Co-Authors

Aïcha Bouzid

Tunis El Manar University

Zied Lachiri

National Engineering School of Tunis

Kaïs Ouni

National Engineering School of Tunis

Dorra Ben Ayed

Hopital Universitaire Habib Bourguiba

Zied Hajaiej

Tunis El Manar University

Rimah Amami

Imam Abdulrahman Bin Faisal University

Asma Rabaoui

Centrale Marseille

Lamia Bouafif

Tunis El Manar University

Mohamed Anouar Ben Messaoud

Tunis El Manar University

Najet Arous

Tunis El Manar University

Cooperative Institutions

National Engineering School of Tunis

Tunis El Manar University

Tunis University

Sidi Mohamed Ben Abdellah University

École Normale Supérieure - PSL

Mohammed V University

Centre National de la Recherche Scientifique

Agence pour le Développement et la Réhabilitation de la Ville de Fès

University of Sfax

Mohamed I University

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Field

Voicing Classification Based on Speech Spectrum Multi-Scale Product

IFAC Proceedings Volumes (2010)

Mohanmed Anouar Ben Messaoud Aïcha Bouzid Noureddine Ellouze

10.3182/20100712-3-fr-2020.00056

Cite

Citations (0)

Primitive Invariants for Handwritten Arabic Script

Sofiene Haboubi Samia Snoussi Maddouri Noureddine Ellouze

Abstract —The choice of relevant features is very decisive inhandwriting recognition rate. Our aim is to present some usefulstructural and statistical features and see their degree of variability. Inthis paper, we start with a description of the variability of the Arabichandwriting and the way how to reduce it. Three kinds of feature setsused by our handwriting systems are then presented evaluated anddiscussed. The comparison is carried on a database of images fromIFN/ENIT databases. The Neural Network Multilayer perceptrons isour method of classiﬁcation. A contrastive study of these primitivesis done according to recognition their time and memory consumingand their variability degree. Keywords —Arabic script, invariant descriptors. I. I NTRODUCTION T HE choice of primitives is very crucial for the hand-writing recognition step. The main purpose is to andan optimal set of features that would properly partition thespace and achieve classiﬁcation. A compromise should berespected during primitive extraction: The extractor shouldprovide uniformed primitives for different kinds of scripts.These primitives should describe all details necessary forrecognition process. Further, it should not be time or memoryspace consuming [1]. Several structural primitives extractionmethods are presented in literature as [2], [3]. These primitivesare based on three principal methods: image projection [4], [5],skeletization [6]–[8] and contour function [9]–[12]. Projectionmethod is the simplest and the least time and memory spaceconsuming for images with a little size but it is very sensitiveto slant and it can not give a precise description of theimage. Skeletization is generally used for script segmentationin uniform graphemes. Cursive Script segmentation difﬁcultiesreduce the efﬁciency of skeleton description. The contourwhich globally describes the image is easy to be detected.By ulterior processing steps we can extract different kind ofstructural, analytic, global or local primitives from the imagecontour. This kind of description is most used in our proposedstructural and analytic or statistical descriptors. HandwritingArabic words recognition becomes a complex task because ofthe diversities of script styles on the one hand and the greatquantities of information given by the Arabic word on theother hand. In this paper we present in section 2 a state of theart on the main useful primitives for handwritten Arabic script.In section 3 we describe three primitives’ extraction methods.In section 4 we evaluate and compare the extraction rate ofeach method on IFN/ENIT database [13].

Cursive

Handwriting

Perceptron

Source

Cite

Citations (0)

Voicing Detection in Noisy Speech Signal

Lecture notes in computer science (2008)

Aïcha Bouzid Noureddine Ellouze

SIGNAL (programming language)

10.1007/978-3-540-69905-7_62

Cite

Citations (0)

Spectral Analysis of DNA Sequence: The Exon's Location Method

Afef Elloumi Oueslati Zied Lachiri Noureddine Ellouze

The system's analysis concepts based on signal processing methods are considered as very important tools in the genomic field's exploration. Spectral analysis of DNA sequences reveals genome's specific periodicities. Studying one organism's genome requires three steps. First, the DNA coding method: the DNA's string data are transformed into numerical signal. Second periodicities are detected by spectral analysis. Third, following the evolution of this periodicity along the genome allows the exon's, or protein coding region's, location. In this paper, we expose the exon's location method: including the coding and spectral analysis steps. After, we propose a two dimensional graphical representation for the spectrum. The experimental result including a variety of coding indicator's methods reveals the importance of each base's combination, from one base to four group bases, in the exon's enhancement. The contributions in percent of each of this group's indicator are finally estimated.

Coding region

genomic DNA

10.1109/icdsp.2007.4288532

Cite

Citations (6)

Robustness Improvement of Speaker Segmentation techniques Based on the Bayesian Information Criterion

Hachem Kadri Zied Lachiri Noureddine Ellouze

Speaker segmentation is the problem of finding speaker segment boundaries when a speaker begins and stop speaking in an audio speaker stream. This segmentation of audio data is of interest to a broad class of applications like surveillance meetings summarization or indexing of broadcast news. Unsupervised speaker segmentation approaches suppose that there is no information about the speakers and their number is known a priori. It can be classed into three categories: energy-based segmentation, metric-based selection and model-selection-based segmentation. Energy-based segmentation: silence in the input audio stream is detected either by a decoder or directly by measuring and thresholding the audio energy. The segments are then generated by cutting the input at silence locations. • Metric-based segmentation: the audio stream is segmented at maxima of the distances between neighboring windows placed in evenly spaced time intervals. • Model-selection-based segmentation [4]: assuming that data are generated by a Gaussian process, speaker changes are detected by using a statistical decision criterion within a sliding window through the audio stream. A widely used technique for speaker segmentation is based on the Bayesian Information Criterion (BIC). Indeed, BIC segmentation presents the advantages of robustness and threshold independence. However, this method, extremely computationally expensive, can introduce an estimation error due to insufficient data when the speaker turns are close to each other.

Bayesian information criterion

Robustness

Speech segmentation

Segmentation-based object categorization

10.1109/ictta.2006.1684567

Cite

Citations (1)

On the Search of Organization Measures for a Kohonen Map Case Study: Speech Signal Recognition

International Journal of Digital Content Technology and its Applications (2010)

Najet Arous Noureddine Ellouze

Unsupervised learning scheme like the self-organizing map (SOM) has been used to classify speech sounds in an ordered manner. SOM is able to extract the most salient features of the input signal and provides a simple way of visualizing them. The distance between two units on the map was used as an objective measure of their perceptual similarity. This paper presents a study of the evaluation of a SOM trained by a sequential learning algorithm integrating information enrichment principal. Two complementary analyses are proposed: quantitative analysis and qualitative one. SOM has been used to visualize speech database as a phenotypic map. The latter was used to generate quantitative measures of the input space. In this paper, we propose different organization measures of a Kohonen map. Some of them are restricted to evaluate a map without referring to data manifold. Others are restricted to quantify map organization with respect to data manifold. We propose also how these organization measures act as fitness measures. These organizations measures are evaluated on the case study of phoneme classification of TIMIT acoustic-phonetic continuous speech corpus. The experimental results show that the proposed combined organization measures provide more significant values according to map sizes.

Self-organizing map

SIGNAL (programming language)

10.4156/jdcta.vol4.issue3.7

Cite

Citations (10)

Comparative analysis between a variational method and wavelet method PURE-LET to remove poisson noise corrupting CT images

2022 19th International Multi-Conference on Systems, Signals & Devices (SSD) (2016)

Kais Ghazouani Noureddine Ellouze Fehry Mohamed Chedly

Poisson noise removal has been implemented by several methods and several approaches. This type of noise generally affects several types of images and in the whole case medical images especially those rebuilt after an X radiation. About it, as tests images are CT ones, further, we will present two methods derived from two different approaches to Poisson noise removal. one that is variational, which is a restoring process that takes into account the actual appearance of this noise which is multiplicative, the other methods derived from wavelet approach which is among the growing methods for the Poisson noise removal mainly in medical dataset. This method is called PURE-LET, simply to say Poisson Unbiased Risk Estimation-Linear Expansion of thresholds. This latter is based on of an optimization of statistical tool Mean Square Error (MSE) using an unnormalized Haar decomposition wavelet transform (DWT) and exploiting the property of the orthogonality of these latter. An original comparison at the end of this paper allows us to assess the reliability and robustness of either the variational method or the PURE-LET one against the annoying factors in medical imaging such as artifact, spatial resolution, etc. Finally, we come to the conclusion that the Variational method excels at the Pure-Let wavelet method.

Robustness

Variational method

Orthogonality

10.1109/ssd.2016.7473733

Cite

Citations (3)

Speech classification in noisy environment using subband decomposition

Zied Lachiri Noureddine Ellouze

This paper presents a new algorithm for robust speech classification in adverse conditions, using an appropriate wavelet packet decomposition of the speech signal. The classification is achieved by generating a correlation functions of different subbands signals derived from a tree structured filter banks. The performance of the proposed technique is evaluated on speech signal (timit database) with real world noise, added to it, at various SNR. Experimental results show the accuracy of the proposed technique especially in low signal to noise circumstance (/spl les/ 10 dB).

TIMIT

Filter bank

SIGNAL (programming language)

10.1109/isspa.2003.1224727

Cite

Citations (6)

Asr systems in noisy environment: Auditory features based on gammachirp filter using the AURORA database

Hajer Rahali Zied Hajaiej Noureddine Ellouze

Cite

Citations (0)

Table of contents

Abir Zaidi Abdennaceur Baghdad Abdelhakim Ballouk Abdelmajid Badri Wahaj Abbas Awan