Browsing by Author "Supervisor: ASSAS, Ouarda"

Now showing 1 - 4 of 4
  • Thumbnail Image
    ItemOpen Access
    Design and Realization of Multimedia Data Encryption System
    (University of M'sila, 2016-06-10) GUESMIA, Seyf Eddine; Supervisor: BOUDERAH, Brahim; Supervisor: ASSAS, Ouarda
    The Cryptology is the most indispensabte science used in the guarantee of the confidentiality of the exchanged information. It consist of creating and studying algorithms that secure the transnission over the intemet. A new algorithrn appeared in 2011, similar to the ariginal AES algorithm, provides more security by enhancing the reliability against the brute force attack. How6ver, it cannot handle the heavy encryption-processing load with the existence of the modern ressurce-limited systems. Our work presents an improvement design of 512-bit AES *lgorithm that provides high security level and ameliorates the prfonnance by minimizing the use of memory space and time eacryption to be able to work in specific characteristics of resourceJ imited systems
  • Thumbnail Image
    ItemOpen Access
    Identifying Individuals using Robust Iris Recognition Techniques
    (University of M'sila, 2015-06-10) ZEMOURI, Khadidja; Supervisor: ASSAS, Ouarda; Supervisor: BENOUIS, Mohamed
    The increasing need for information security has led to more attention being given to biometrics-based, automated personal identification. Among existing biometric approaches, the human iris is the most promising technique. This work is proposed the Iris recognition method using local binary pattern(LBP) and component analysis(PCA) with measure distance. In addition, we have introduced the multi biometrics system particularly multimodal system to improve the performance of the identification system. For the validation of this work, we use database CASIA VI. Experimental results show that the multimodal fusion scenario (Iris + Face) with the application of LBP and LDA on the iris and face respectively gives the best recognition rate of 98.33% using mean rule of fusion.
  • Thumbnail Image
    ItemOpen Access
    Postal Code Handwritten Recognition System
    (University of M'sila, 2016-06-10) Amroune, Abdelheq; Supervisor: ASSAS, Ouarda; Supervisor: Fernini, L. Belabdelouahab
    A three based classifiers system was created. A back-propagation neural network With one hidden layer , a support Vector machine classifier, and K- nearest Neighbor classifier were used to create an adaptive postal code digits recognition system by using the Hu moments invariants feature extraction method. The system was trained and evaluated through different forms of handwriting samples provided by both male and female participants. Experiments tested, the effect of the size set on the recognition accuracy, and the effect of handwriting style on the recognition accuracy. Results showed that the handwriting style of the subjects had varying and drastic effects on the recognition accuracy which allowed to identify some of the problems With the system digits encoding.
  • Thumbnail Image
    ItemOpen Access
    vehicle Identification by license plate using support vector machine (SVM)
    (University of M'sila, 2016-06-10) Boukhercha, Ryma; Supervisor: ASSAS, Ouarda
    La reconnaissance automatique de plaques immatriculation est une méthode de surveillance de masse qui utilise la reconnaissance optique de caracteres sur les images pour lire les plaques d'immatriculation des véhicules, il _joue un röle important dans diverses applications liées au systeme de transport automatisé telles que la surveillance de la circulation routiérey la détection de véhicules volés, paiements automatiques de péages sur les routes ou les ponts, les pares de stationnement contröle d'accés, etc. Il doit étre un processus des plaques immatriculation (LP) rapidement et avec succes dans différentes conditions environnementales, comme å l'intérieur, extérieur, jour ou de nuit. Dans ce travau nous montrons différentes techniques utilisées pour ANPR, utilisé deux méthodes: moment de Hu et le modéle locale binaire (LBP) pour Itextraction de caractéristiques, et deux méthodes pour la classification : machine vecteurs de support (SVM) et le modéle correspondant (KNN). Dans les résuitats expérimentaux, nous a obtenu le meilleur taux de reconnaissance de et SVM