Khalid Saeed

Region growing based segmentation algorithm for typewritten and handwritten text recognition

This paper presents a new technique of high accuracy to recognize both typewritten and handwritten English and Arabic texts without thinning. After segmenting the text into lines (horizontal segmentation) and the lines into words, it separates the word into its letters. Separating a text line (row) into words and a word into letters is performed by using the region growing technique (implicit segmentation) on the basis of three essential lines in a text row. This saves time as there is no need to skeletonize or to physically isolate letters from the tested word whilst the input data involves only the basic information-the scanned text. The baseline is detected, the word contour is defined and the word is implicitly segmented into its letters according to a novel algorithm described in the paper. The extracted letter with its dots is used as one unit in the system of recognition. It is resized into a 9x9 matrix following bilinear interpolation after applying a lowpass filter to reduce aliasing. Then the elements are scaled to the interval [0,1]. The resulting array is considered as the input to the designed neural network. For typewritten texts, three types of Arabic letter fonts are used-Arial, Arabic Transparent and Simplified Arabic. The results showed an average recognition success rate of 93% for Arabic typewriting. This segmentation approach has also found its application in handwritten text where words are classified with a relatively high recognition rate for both Arabic and English languages. The experiments were performed in MATLAB and have shown promising results that can be a good base for further analysis and considerations of Arabic and other cursive language text recognition as well as English handwritten texts. For English handwritten classification, a success rate of about 80% in average was achieved while for Arabic handwritten text, the algorithm performance was successful in about 90%. The recent results have shown increasing success for both Arabic and English texts.

Cursive-Character Script Recognition Using Toeplitz Model and Neural Networks

This paper presents a hybrid method to use both the idea of projection and Toeplitz Matrix approaches to describe the feature vectors of an image and hence identifying it. The method applies two different tools. The main one is Toeplitz forms and the second is Neural Networks. The image model considered in this work are some selected Arabic scripts. The letter is first projected on 12 axes, then the lengths of these axes are measured and afterwards for the sake of classification and recognition these lengths are compared with the ones in the data base. The method has proved its high efficiency upon the other known approaches. Toeplitz model has shown its successful role in improving the description of the image feature vectors and hence increasing the rate of recognition. The overall algorithm has reached a very low rate of misclassification. Both machine and hand written cases have been studied. In this paper, examples of handwritten scripts are considered.

Experimental Algorithm for Characteristic Points Evaluation in Static Images of Signatures

The paper presents experimental method for the extraction of handwritten signature features with the aim of incorporating them in the offline signature recognition system. The algorithm uses view-based approach and searches for the extreme values with the threshold value being applied. This investigation is a continuation of previous work extended with experiments on classification of resulted feature vectors. The classification of feature vectors is conducted by means of Dynamic Time Warping (DTW) algorithm. Experiments were carried out with the standard DTW algorithm with window and slope constraints.

A New Step in Arabic Speech Identification: Spoken Digit Recognition

This work presents a new Algorithm to recognize separate voices of some Arabic words, the digits form zero to ten. Firstly we prepare our signal by pre-processing trial. Next the speech signal is processed as an image by Power Spectrum Estimation. For feature extraction, transformation and hence recognition, the algorithm of minimal eigenvalues of Toeplitz matrices together with other methods of speech processing and recognition are used. At the stage of classification many methods are tested from classical ones, which depend on the matrix theory, to different types of neuron networks, mainly radial basis functions neural networks. The success rate obtained in the presented experiments is almost ideal and exceeded 98% for many cases. The results have shown flexibility to extend the algorithm to speaker identification.

Modular logic of authentication using dynamic keystroke pattern analysis

Authenticating users in a continual manner has become extremely critical for a wide range of applications in the domain of pervasive computing and Internet of Things (IoT). In these days, it’s also an accepted fact that user authentication based on biometric features is often more efficient than the traditional means of password-based authentication. However, many of the existing biometric techniques like Iris or finger-print recognition are effective only when the person to be authenticated or verified is physically accessible. Thus such technologies are good for applications like Passport Control and fall short of the requirements for IoT applications like an integrated remote-healthcare where different types of users like Doctors, patients, hospitals, insurance companies, other care-givers and even authorized civic-body administrators are to be continually authenticated from remote locations. It is important to ensure that the desired services are accessed only by a legitimate user and no one else. In ...

New Experiments on Word Recognition Without Segmentation

A new hybrid system for word recognition is discussed in this work. The system is based on a modification to the view-based approach presented in authors’ previous works. The system does not need thinning or segmentation of the analyzed word. The word is treated as a whole image. The characteristic vectors taken from both top and bottom views of the image are processed with the method of minimal eigenvalues of Toeplitz matrices. The obtained series of minimal eigenvalues are used for classification with Artificial Neural Networks. The results of the experiments on a set of common English words are presented.

Intelligent Feature Extract System for Cursive-Script Recognition

The paper describes a newly presented hybrid method for high efficiency in script image feature extraction. The recognition rate was about 82% for very large number of scripts per class. However, it has reached even 100% in some cases with a smaller number of scripts per class. The system contains two projection-based methods for image characteristics extraction presented by very simple feature vectors and one image descriptor. A specially worked out thinning algorithm for the recognition system has simplified the feature extracting procedure as it provides a continuous one-pixel width skeleton of the script, which is essential for the simple-projection approach.

A Study on the Importance of Biometric Technique Selection in the Protection of Company Resources

In this paper some modern methods of physiological identification of people, the importance of biometrics in company management, and technical resources used in this field are presented. Authors describe the achievements of biometrics and its efficiency in ensuring the security of data and equipment resources from industrial management point of view. Also given the advantages of the solutions that are based on biometric methods, and evidence showing their increasing significance in company activity and its development.

The impact of database quality on keystroke dynamics authentication

This paper concerns keystroke dynamics, also partially in the context of touchscreen devices. The authors concentrate on the impact of database quality and propose their algorithm to test database quality issues. The algorithm is used on their own as well as the well-known . Following specific problems were researched: classification accuracy, development of user typing proficiency, time precision during sample acquisition, representativeness of training set, sample length.

Signature verification by only single genuine sample in offline and online systems

The paper presents innovatory methods and algorithms with experimental results on signature verification. It is mainly focused on applications where there is only one reference signature available for comparison. Such restriction is often present in practice and requires selection of specific methods. In this context, both offline and online approaches are investigated. In offline approach, binary image of the signature is initially thinned to obtain a one pixel-wide line. Then, a sampling technique is applied in order to form the signature feature vector. The identification and verification processes are based on comparing the reference feature vector with the questioned samples using Shape Context algorithm. In the case of online data, the system makes use of dynamic information such as trajectory, pen pressure, pen azimuth and pen altitude collected at the time of signing. After further preprocessing, these functional features are verified by means of Dynamic Time Warping method.