Najoua Essoukri Ben Amara

Planar Markov modeling for Arabic writing recognition: advancement state

In this paper, we show how planar hidden Markov models (PHMM) can offer great potential to solve difficult Arabic character recognition problems, especially its cursivness. A convenient architecture is defined for printed Arabic sub-words. It yields an easy solution to implement the modeling of the different morphological variations of the Arabic writing, i.e., vertical and variable horizontal linkages. A more flexible architecture, developed for Arabic handwritten words, is under test. The structure proposed presents the aptitude to absorb the variability of the manuscript. Indeed, the experiments have shown promising results and directions for further improvements. In the present paper, we describe both retained architectures, showing the applicability of the PHMM to the Arabic complexities. This is owed precisely to the definition of the PHMM, which permits to follow efficiently the natural variations in bands of the Arabic script.

Multifont Arabic character recognition using Hough transform and hidden Markov models

Optical characters recognition (OCR) has been an active subject of research since the early days of Computers. Despite the age of the subject, it remains one of the most challenging and exciting areas of research in computer science. In recent years it has grown into a mature discipline, producing a huge body of work. Arabic character recognition has been one of the last major languages to receive attention. This is due, in part, to the cursive nature of the task since even printed Arabic characters are in cursive form. This paper describes the performance of combining Hough transform and hidden Markov models in a multifont Arabic OCR system. Experimental tests have been carried out on a set of 85,000 samples of characters corresponding to 5 different fonts from the most commonly used in Arabic writing. Some promising experimental results are reported.

Unsupervised facial expressions recognition and avatar reconstruction from kinect

This paper presents a solution capable of recognizing the facial expressions performed by a persons face and mapping them to a 3D face virtual model using the depth and RGB data captured from Microsofts Kinect sensor. This solution starts by detecting the face and segmenting its regions, then, it identifies the actual expression using EigenFaces metrics on the RGB images and reconstructs the face from the filtered Depth data. A new dataset relative to 20 human subjects is introduced for learning purposes. It contains the images and point clouds for the different facial expressions performed. The algorithm seeks and displays automatically the seven state of the art expressions including surprise, fear, disgust, anger, joy, sadness and the neutral appearance. As result our system shows a morphing sequence between the sets of 3D face avatar models.

A dataset for Arabic text detection, tracking and recognition in news videos- AcTiV

Recently, promising results have been reported on video text detection and recognition. Most of the proposed methods are tested on private datasets with non-uniform evaluation metrics. We report here on the development of a publicly accessible annotated video dataset designed to assess the performance of different artificial Arabic text detection, tracking and recognition systems. The dataset includes 80 videos (more than 850,000 frames) collected from 4 different Arabic news channels. An attempt was made to ensure maximum diversities of the textual content in terms of size, position and background. This data is accompanied by detailed annotations for each textbox. We also present a region-based text detection approach in addition to a set of evaluation protocols on which the performance of different systems can be measured.

Dynamic hierarchical Bayesian network for Arabic handwritten word recognition

This paper presents a new probabilistic graphical model used to model and recognize words representing the names of Tunisian cities. In fact, this work is based on a dynamic hierarchical Bayesian network. The aim is to find the best model of Arabic handwriting to reduce the complexity of the recognition process by permitting the partial recognition. Actually, we propose a segmentation of the word based on smoothing the vertical histogram projection using different width values to reduce the error of segmentation. Then, we extract the characteristics of each cell using the Zernike and HU moments, which are invariant to rotation, translation and scaling. Our approach is tested using the IFN / ENIT database, and the experiment results are very promising.

Artificial human face recognition via Daubechies wavelet transform and SVM

This work presents an approach for applying face recognition to non-biological entities (avatars) in virtual worlds to achieve authentication. Massively multiplayer online games involve virtual worlds which require avatar identification to avoid fraud. First, virtual worlds and avatars are briefly discussed. Next, the concepts of facial biometrics and the face recognition systems are presented. Later, support vector machines and wavelet transforms are introduced as classification tools. Finally, the dataset and the designed biometric system are described with the obtained results.

Bimodal biometric verification with different fusion levels

In this paper, we propose the fusion of two unimodal biometric verification systems, based on off-line signature and handwriting. The fusion of modalities is made at four different levels. In the first case, the fusion is situated at the feature extraction level by concatenating the vectors relative to each modality. In the second case, the fusion is situated at the decision level by the logic operators ‘AND’ and ‘OR’. Fusion at the matching score level is realised with two methods: arithmetic average and support vector machines. Fusing information from biometric images is based on a discrete wavelet transform. According to 12000 samples of handwritings and signatures, the best results are obtained for the feature extraction level and the matching score level.

Evaluation of SVM classification of avatar facial recognition

The security of virtual worlds has become a major challenge. The huge progress of Internet technologies, the massive revolution of media and the use of electronic finance increasingly deployed in a world where the law of competition forces financial institutions to devote huge amounts of capital to invest in persistent digital worlds like Second Life or Entropia Universe whose economic impact is quite real [1]. Thus, virtual communities are rapidly becoming the next frontier for the cyber-crime. So, it is necessary to develop equitable tools for the protection of virtual environments, similar to those deployed in the real world, such as biometric security systems. In this paper, we present a biometric recognition system of non-biological entities (avatar) faces based on exploration of wavelet transform for characterization and Support Vector Machines for classification. This system is able to identify and verify avatars during their access to certain information or system resources in virtual communities. We also evaluate the performance of our avatar authentication approach focusing specifically on the classification stage. Obtained results are promising and encouraging for such first contribution within the framework of this new research field.

Face recognition improvement using soft biometrics

In this work we are interested in soft facial biometrics, a new field that aims at strengthening the performance of primary biometric systems based on traditional ways of biological type (DNA, saliva), morphological (face, iris, fingerprint) or behavioral (signature, handwriting, voice). We propose three types of facial soft biometrics: facial measurements, skin color and hair color. The results show that the fusion of these modalities with the primary face recognition system based on wavelet characterization and SVM training has increased the recognition rate and has decreased the equal error rate.

Development of a database with ground truth for old documents analysis

Standard databases play essential roles for evaluating and comparing results obtained by different groups of researchers. In this paper, a database of images derived from the Tunisian heritage is introduced. We present first the importance of digitization for the preservation and the protection of patrimonial collections. Then, we describe the development process of our database, intended for the analysis of historical document images, in collaboration with the National Archives of Tunisia. In the second place, we demonstrate the importance of the ground truth for image databases and the annotation impact of these images. Thus, we present our approach for images annotation which consists in producing manually and collectively keywords for each image.