Multi-character Field Recognition for Arabic and Chinese Handwriting

A system for writer identification based on Arabic handwritten words was built. First a database of words was gathered and used as a test base. Then, features vectors were extracted from writers' word images. Prior to feature extraction, normalization operations were applied to a word or text line. In this research, we studied the feature extraction and recognition operations on Arabic text, on the identification rate of writers. Since there is no well known database containing Arabic handwritten words for researchers to test, we built a new database of off-line Arabic handwriting text to be used for writer identification research. The proposed database is meant to provide training and testing sets for Arabic writer identification research. Arabic handwritten words were collected from 100 writers. We evaluated the performance of edge-based directional probability distributions as features and other features in Arabic writer identification.

This paper presents the use of Hidden Markov Models (HMM) to classifying Arabic words that have been represented by moments into a set of predefined classes. The proposed scheme results in enhanced performance because it deals effectively with errors that commonly originate during the feature extraction stage and it accounts for variations due to the individual handwriting performance of different data sets. To overcome the variations in images, each is represented by rotation and scale invariant feature vector known as Pseudo Zernike Moments (PZM). To demonstrate its effectiveness, the proposed classification scheme has been employed in the context of different sets of handwritten Arabic words. The classification results demonstrate that the proposed approach can classify handwritten Arabic word with a error rate less than 10%.

Offline Arabic handwritten word recognition: A transfer learning approach

Erasing Chinese handwritten words in images has many applications in our daily life. It can hide the private information in the archive images (home address, identity information), delete the signature information in the contract images, and clear the handwriting in the test paper to reuse it. However, the existing research mainly focuses on scene text removal. To fill this gap and facilitate this research direction, we propose a novel model called CHENet (Chinese Handwriting Eraser Network) that can automatically remove Chinese handwriting located on the images. In particular, we design two modules: CAB (Convolution Attention Block) and CAP (Chain Attention Pooling) that can improve the ability of the CHENet. Also, for information protection, we collect images with Chinese handwritten words from the archive, contract, and exam papers, and build the CH-dataset that consists of 1623 scrupulously annotated images. Qualitative and quantitative evaluations show that our CHENet can solve this task well.KeywordsText removalChinese handwriting removalEnd-to-end networkAttention network

In this work, we investigate the combination of PGM (Propabilistic Graphical Models) classifiers, either independent or coupled, for the recognition of Arabic handwritten words. The independent classifiers are vertical and horizontal HMMs (Hidden Markov Models) whose observable outputs are features extracted from the image columns and the image rows respectively. The coupled classifiers associate the vertical and horizontal observation streams into a single DBN (Dynamic Bayesian Network). A novel method to extract word baseline and a simple and easily extractable features to construct feature vectors for words in the vocabulary are proposed. Some of these features are statistical, based on pixel distributions and local pixel configurations. Others are structural, based on the presence of ascenders, descenders, loops and diacritic points. Experiments on handwritten Arabic words from IFN/ENIT strongly support the feasibility of the proposed approach. The recognition rates achieve 90.42% with vertical and horizontal HMM, 85.03% and 85.21% with respectively a first and a second DBN which outperform results of some works based on PGMs.

This paper presents an overview of feature extraction techniques for unconstrained Arabic handwritten word recognition. Choosing a technique for extraction the features considers the most important factor in achieving high recognition rates in word or character recognition. Different techniques were designed to extract the features from the Arabic words. These techniques are presented and discussed in terms of invariant invariance properties.

Arabic handwritten text recognition is a complex and challenging research domain. This study proposes an offline Arabic handwritten word recognition system based on transfer learning. The system exploits four pre-trained convolutional neural network (CNN) architectures, namely VGG16, ResNet50, AlexNet, and InceptionV3. In addition, a specialized image recognition model derived from the ImageNet dataset is incorporated. A combination strategy is designed to combine transfer learning with specific fine-tuning techniques, aiming to improve recognition accuracy. The study is conducted on the IFN/ENIT dataset, which includes images of Tunisian City and village names. The results show that the proposed system achieves a recognition accuracy of 94.73%, which is significantly higher than the accuracy rates achieved by previous approaches. These results suggest that the proposed system is a promising approach for Arabic handwritten text recognition.

Hidden Markov Models (HMMs) are now widely used in off-line handwriting recognition and, in particular, in Arabic handwritten word recognition. In contrast to the conventional approach, based on Gaussian mixture HMMs, we have recently proposed to directly fed columns of raw, binary pixels into Bernoulli mixture HMMs. In this work, column bit vectors are extended by means of a sliding window of adequate width to better capture image context at each horizontal position of the word image. Using these windowed Bernoulli mixture HMMs, very good results are reported on the well-known IfN/ENIT database of Arabic handwritten Tunisian town names.

Lacking of dataset is still a serious problem for researchers who study on online handwriting word recognition (HWR). In this paper, a handwritten Chinese word synthesis method is proposed for the first time to generate a large scale handwritten Chinese word dataset. The distributions of shape and position characteristics, such as aspect radio, character interval and the angle of gravity center line in each word sample of the Word8888 dataset have been estimated respectively. Based on this, we synthesize as large as 44,208 categories of 8,311,104 unconstrained handwritten Chinese word samples. To verify the validity of the synthesized dataset, a practical rotation free handwriting Chinese word recognition system is presented based on a new holistic approach. Experimental results for randomly rotated word samples demonstrate that the holistic approach can achieve 91.96% recognition accuracy, which provides evidence for the effectiveness of our method.

This paper proposes a novel similarity measure between vector sequences. We work in the framework of model-based approaches, where each sequence is first mapped to a Hidden Markov Model (HMM) and then a measure of similarity is computed between the HMMs. We propose to model sequences with semicontinuous HMMs (SC-HMMs). This is a particular type of HMM whose emission probabilities in each state are mixtures of shared Gaussians. This crucial constraint provides two major benefits. First, the a priori information contained in the common set of Gaussians leads to a more accurate estimate of the HMM parameters. Second, the computation of a similarity between two SC-HMMs can be simplified to a Dynamic Time Warping (DTW) between their mixture weight vectors, which significantly reduces the computational cost. Experiments are carried out on a handwritten word retrieval task in three different datasets-an in-house dataset of real handwritten letters, the George Washington dataset, and the IFN/ENIT dataset of Arabic handwritten words. These experiments show that the proposed similarity outperforms the traditional DTW between the original sequences, and the model-based approach which uses ordinary continuous HMMs. We also show that this increase in accuracy can be traded against a significant reduction of the computational cost.

The application of sentiment analysis, also known as opinion mining, is more difficult in Chinese than in Indo-European languages, due to the compounding nature of Chinese words and phrases, and relatively lack of reliable resources in Chinese. This study used seed words, Chinese morphemes, which are mono-syllabic characters that function as individual words or be combined to create Chinese words and phrases, to classify movie reviews found on Yahoo! Taiwan. We utilized higher Pointwise Mutual Information (PMI) collocations, which consist of selected morpheme-level compounded features to build classifiers. The contributions of this study include the following: (Bird 2006) proposing a method of generating domain-dependent Chinese morphemes directly from large data set without any predefined sentimental resources; (Bradley and Lang 1999) building morpheme-based classifiers applicable in various movie genres, and shown to produce better results than other classifiers based on keywords (NTUSD and HowNet) or feature selection (TFIDF); (Church and Hanks in Computational linguistics, 16(1), 22-29 1990) identifying compounds that have different semantic polarities depending on contexts.

In this paper, we present the first truly 2-D recognition system of Arabic handwritten words. The used recognizer is the NSHPZ-HMM model with a new zoning design approach. The zoning method we propose depends on the baseline of the treated word. The key point relies in the ability of our baseline to perfectly follow the writing curve even in case of vertical ligatures and of different slant angles within the same Arabic word. The result is an appropriate division in zones of an image word that facilitates its modelization through the use of the NSHPZ-HMM model. Experiments on Arabic literal amounts of AHDB database show the efficiency of the proposed recognition system over an NSHPZ-HMM using zoning independent baseline, a classical pixel-based NSHP-HMM and a set of state of the art systems developed on the same database.

We propose C-SAR, a Class-specific and Adaptive Recognition algorithm for Arabic handwritten Cheques. Existing methods suffer from low accuracy due to the complex structure of Arabic script and high-dimensional datasets. In this paper, we present an adaptive algorithm that implements a class-specific classification to address these challenging issues. C-SAR trains a set of class-specific machine learning models of Support Vector Machines and Artificial Neural Networks features extracted using angular pixel distribution approach. Furthermore, we propose a class-specific taxonomy of Arabic cheque handwritten words. The proposed taxonomy divides the Arabic words into groups over three layers based on their structural characteristics. Accordingly, C-SAR performs classification on three phases, i.e., 1) similar and non-similar structures, for binary classification, 2) classes with similar structures into another two categories, and 3) class-specific models to recognize the Arabic word from the given image. We introduce benchmark experimental results of our method against previous methods on the Arabic Handwriting Database for Text Recognition. Our method outperforms the baseline methods with at least 5% accuracy having 90% average classification accuracy.KeywordsHandwritten recognitionImage classification

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.

<span lang="EN-US">Arabic handwritten text recognition has long been a difficult subject, owing to the similarity of its characters and the wide range of writing styles. However, due to the intricacy of Arabic handwriting morphology, solving the challenge of cursive handwriting recognition remains difficult. In this paper, we propose a new efficient based image processing approach that combines three image descriptors for the feature extraction phase. To prepare the training and testing datasets, we applied a series of preprocessing techniques to 100 classes selected from the handwritten Arabic database of the </span><em><span lang="EN-US">Institut Für Nachrichtentechnik/Ecole Nationale d'Ingénieurs de Tunis</span></em><span lang="EN-US"> (IFN/ENIT). Then, we trained the k-nearest neighbor’s algorithm (k-NN) algorithm to generate the best model for each feature extraction descriptor. The best k-NN model, according to common performance evaluation metrics, is used to classify Arabic handwritten images according to their classes. Based on the performance evaluation results of the three k-NN generated models, the majority-voting algorithm is used to combine the prediction results. A high recognition rate of up to 99.88% is achieved, far exceeding the state-of-the-art results using the IFN/ENIT dataset. The obtained results highlight the reliability of the proposed system for the recognition of handwritten Arabic words.</span>