Anis Mezghani

A Database for Arabic Handwritten Text Image Recognition and Writer Identification

Standard databases play essential roles for evaluating and comparing results obtained by different groups of researchers. In this paper, an Arabic Handwritten Text Images Database written by Multiple Writers (AHTID/MW) is introduced. This database can be used for research in the recognition of Arabic handwritten text with open vocabulary, word segmentation and writer identification. The AHTID/MW contains 3710 text lines and 22896 words written by 53 native writers of Arabic. In addition, ground truth annotation is provided for each text image. The database is freely available for worldwide researchers.

ICFHR2014 Competition on Arabic Writer Identification Using AHTID/MW and KHATT Databases

This paper describes the first edition of the Arabic writer identification competition using AHTID/MW and KHATT databases held in the context of the 14th International Conference on Frontiers in Handwriting Recognition (ICFHR2014). This competition has used the new freely available Arabic Handwritten Text Images Database written by Multiple Writers (AHTID/MW) and the Arabic handwritten text database called KHATT presented in ICFHR2012. We propose three tasks in this Arabic writer identification competition: the first and second are based respectively on word and text line level using the AHTID/MW database and the third one is paragraph based using the KHATT database. We received one system for the second task, three systems for the third task and none for the first task. All systems are tested in a blind manner using a set of images kept internal. A short description of the participating groups, their systems, the experimental setup, and the observed results are presented.

Printed/handwritten Arabic script identification using local features and GMMs

Since printed/handwritten Arabic text recognition is a very challenging research field and the recognition methodologies are different, it is important to separate these two types of texts before the recognition phase. In this paper, we introduce a simple and effective method to identify printed and handwritten Arabic words using local features. A Gaussian Mixture Models (GMMs) based approach is used to model the printed and handwritten classes. Experimental results using some parts of the freely available IFN/ENIT, AHTID/MW and APTI databases show that our method is robust and provides very good identification performance.

Window-based feature extraction framework for machine-printed/handwritten and Arabic/Latin text discrimination

In this paper, we propose a new writing type and script text classification technique to recognize the identity of texts extracted from heterogeneous document images. English, French and Arabic languages are used in these documents with mixed handwritten and machine-printed types. In order to identify each text-line/word image, we propose to use 23 features computed on a fixed-length sliding window. Gaussian Mixture Models (GMMs) are used to achieve the classification objective. This framework has been tested on machine-printed and handwritten text-blocks, text-lines and words extracted from different document images of the Maurdor database. Experimental results reveal the effectiveness of our proposed system in writing type and script identification.

Arabic handwritten word recognition with large vocabulary based on explicit segmentation

In this paper, an Arabic handwritten word recognition system for wide vocabulary using the analytical approach is presented. After the segmentation phase, the characters are recognized using assumption generation of letters. Thus, the proposed system generates assumptions of words by concatenation of letters assumption. Thereafter, a filtering of these words assumption is achieved thanks to a language dictionary. The system showed good performances when applied on a 500 word images dataset.

Multiset Canonical Correlation Analysis: Texture Feature Level Fusion of Multiple Descriptors for Intra-modal Palmprint Biometric Recognition

This paper describes a novel intra-modal feature fusion for palmprint recognition based on fusing multiple descriptors to analyze the complex texture pattern. The main contribution lies in the combination of several texture features extracted by the Multi-descriptors, namely: Gabor Filters, Fractal Dimension and Gray Level Concurrence Matrix. This means to their effectiveness to confront the various challenges in terms of scales, position, direction and texture deformation of palmprint in unconstrained environments. The extracted Gabor filter-based texture features from the preprocessed palmprint images to be fused with the Fractal dimension-based-texture features and Gray Level Concurrence Matrix-based texture features using the Multiset Canonical Correlation Analysis method (MCCA). Realized experiments on three benchmark datasets prove that the proposed method surpasses other well-known state of the art methods and produces encouraging recognition rates by reaching 97.45% and 96.93% for the PolyU and IIT-Delhi Palmprint datasets.

Writing type, script and language identification in heterogeneous documents

In this paper, we propose a writing type, script and language text classification method to automatically determine the identity of texts segmented from heterogeneous document images. These documents are written in Arabic, French and English languages with mixed machine-printed and handwritten text. To handle such a problem, we treat each text-line/word image with a fixed-length sliding window. Each window is represented with 23 simple and efficient features to achieve the writing type and the script identification goal using Gaussian mixture models (GMM). The proposed approach for language identification is based on a bi-gram analysis of an optical character recognition (OCR) output. Experiments have been conducted with handwritten and machine-printed text-blocks, text-lines and words extracted from the Maurdor database. The results reveal the feasibility of our proposed method in writing type, script and language identification.

Arabic Font Recognition Based on Discret Curvelet Transform

There are many difficulties for Arabic text recognition systems to deal with multi-font and multi-size word/text line images. Some Arabic font families introduce complex variability such as overlaps and ligatures. In this case, developing a cascading system (font recognition followed by font dependent text recognition) has become a necessity. In this paper, we have presented a new font recognition system based on curvelet transform.

A Novel Arabic Writer Identification System Using Texture Feature on Multi-resolution Levels

Recognizing an Arabic text with OCR is a complex task caused by the cursive nature of Arabic script. The Arabic letters change forms not only according to their position in the word, but also according to their font in printed text and to their writer in handwritten text. In fact developing a font recognition system or a writer identification system as a pre-recognition step has become a necessity for Arabic text recognition. In this paper, we present an Arabic script recognition system using Curvelet transform for feature extraction in multi-resolution levels. Also, we used a best feature selection algorithm to increase the feature vector size. To validate our proposed system, we tested our system on Arabic handwriting text database ‘KHATT’ using SVM classification. This experiment show a very interesting results.

Contribution on Character Modelling for Handwritten Arabic Text Recognition

Arabic script is considered to be one of the most complex writing systems, which complicate the text recognition task. Among its complexities, the shape of the character depends according to its position in the word. More than 170 different shapes could be constructed to represent 28 basic letters; some of them are more used than others in the Arabic writing. To make training and recognition of characters more efficient, a study on shape modelling of different handwritten Arabic characters seems to be important. A segmentation-free word recognition system based on Hidden Markov Models (HMMs) is used to conduct this study. Experimental results are given for different sets of shape models using the IFN/ENIT database which contains an important number of handwritten Arabic words covering different writing styles.