George K. Kokkinakis

Automatic text categorization in terms of genre and author

The two main factors that characterize a text are its content and its style, and both can be used as a means of categorization. In this paper we present an approach to text categorization in terms of genre and author for Modern Greek. In contrast to previous stylometric approaches, we attempt to take full advantage of existing natural language processing (NLP) tools. To this end, we propose a set of style markers including analysis-level measures that represent the way in which the input text has been analyzed and capture useful stylistic information without additional cost. We present a set of small-scale but reasonable experiments in text genre detection, author identification, and author verification tasks and show that the proposed method performs better than the most popular distributional lexical measures, i.e., functions of vocabulary richness and frequencies of occurrence of the most frequent words. All the presented experiments are based on unrestricted text downloaded from the World Wide Web without any manual text preprocessing or text sampling. Various performance issues regarding the training set size and the significance of the proposed style markers are discussed. Our system can be used in any application that requires fast and easily adaptable text categorization in terms of stylistically homogeneous categories. Moreover, the procedure of defining analysis-level markers can be followed in order to extract useful stylistic information using existing text processing tools.

Speech enhancement based on audible noise suppression

A novel speech enhancement technique is presented based on the definition of the psychoacoustically derived quantity of audible noise spectrum and its subsequent suppression using optimal nonlinear filtering of the short-time spectral amplitude (STSA) envelope. The filter operates with sparse spectral estimates obtained from the STSA, and, when these parameters are accurately known, significant intelligibility gains, up to 40%, result in the processed speech signal. These parameters can be also estimated from noisy data, resulting into smaller but significant intelligibility gains.

Text genre detection using common word frequencies

In this paper we present a method for detecting the text genre quickly and easily following an approach originally proposed in authorship attribution studies which uses as style markers the frequencies of occurrence of the most frequent words in a training corpus (Burrows, 1992). In contrast to this approach we use the frequencies of occurrence of the most frequent words of the entire written language. Using as testing ground a part of the Wall Street Journal corpus, we show that the most frequent words of the British National Corpus, representing the most frequent words of the written English language, are more reliable discriminators of text genre in comparison to the most frequent words of the training corpus. Moreover, the frequencies of occurrence of the most common punctuation marks play an important role in terms of accurate text categorization as well as when dealing with training data of limited size.

Computer aided diagnosis of breast cancer in digitized mammograms

A computer aided neural network classification of regions of suspicion (ROS) on digitized mammograms is presented in this paper which employs features extracted by a new technique based on independent component analysis. Our approach is concentrated in finding a set of independent source regions that generate the observed mammograms. The coefficients of the linear transformation of the source regions are used as features that describe effectively any normal and abnormal region in digital mammograms as well as benign and malignant ROS in the latter case. Extensive experiments in the MIAS Database have shown a recognition accuracy of 88.23% in the detection of all kinds of abnormalities and 79.31% in the task of distinguishing between benign and malignant regions, outperforming in both cases standard textural features, widely used for cancer detection in mammograms.

Automatic authorship attribution

In this paper we present an approach to automatic authorship attribution dealing with real-world (or unrestricted) text. Our method is based on the computational analysis of the input text using a text-processing tool. Besides the style markes relevant to the output of this tool we also use analysis-dependent style markers, that is, measures that represent the way in which the text has been processed. No word frequency counts, nor other lexically-based measures are taken into account. We show that the proposed set of style markers is able to distinguish texts of various authors of a weekly newspaper using multiple regression. All the experiments we present were performed using real-world text downloaded from the World Wide Web. Our approach is easily trainable and fully-automated requiring no manual text preprocessing nor sampling.

Connection-dependent threshold model: a generalization of the Erlang multiple rate loss model

In this paper first, we review two extensions of the Erlang multi-rate loss model (EMLM), whereby we can assess the call-level quality-of-service (QoS) of ATM networks. The call-level QoS assessment in ATM networks remains an open issue, due to the emerged elastic services. We consider the coexistence of ABR service with QoS guarantee services in a VP link and evaluate the call blocking probability (CBP), based on the EMLM extensions. In the first extension, the retry models, blocked calls can retry with reduced resource requirements and increased arbitrary mean residency requirements. In the second extension, the threshold models, for blocking avoidance, calls can attempt to connect with other than the initial resource and residency requirements which are state dependent. Secondly, we propose the connection-dependent threshold model (CDTM), which resembles the threshold models, but the state dependency is individualized among call-connections. The proposed CDTM not only generalizes the existing threshold models but also covers the EMLM and the retry models by selecting properly the threshold parameters. Thirdly, we provide formulas for CBP calculation that incorporate bandwidth/trunk reservation schemes, whereby we can balance the grade-of-service among the service-classes. Finally, we investigate the effectiveness of the models applicability on ABR service at call set-up. The retry models can hardly model the behavior of ABR service, while the threshold models perform better than the retry models. The CDTM performs much better than the threshold models; therefore we propose it for assessing the call-level performance of ABR service. We evaluate the above-mentioned models by comparing each other according to the resultant CBP in ATM networks. For the models validation, results obtained by the analytical models are compared with simulation results.

An unconstrained handwriting recognition system

Abstract. In this paper, an integrated offline recognition system for unconstrained handwriting is presented. The proposed system consists of seven main modules: skew angle estimation and correction, printed-handwritten text discrimination, line segmentation, slant removing, word segmentation, and character segmentation and recognition, stemming from the implementation of already existing algorithms as well as novel algorithms. This system has been tested on the NIST, IAM-DB, and GRUHD databases and has achieved accuracy that varies from 65.6% to 100% depending on the database and the experiment.

Skew angle estimation for printed and handwritten documents using the Wigner–Ville distribution

A skew estimation algorithm for printed and handwritten documents, based on the document’s horizontal projection profile and its Wigner – Ville distribution, is presented. The proposed algorithm is able to correct skew angles that range between 289 and þ898 detecting the right oriented position of the page by the alternations of the horizontal projection profile. It is able of processing successfully handwritten documents, even if they consist of non-parallel text lines. It deals with the presence of graphics, while a few text lines suffice for the application of the algorithm. Furthermore, the latter permits the use of only a part of the page for the skew estimation minimizing the computational complexity. The proposed algorithm was evaluated on a wide variety of pages (i.e. printed, handwritten, multi-column, application forms etc.) achieving a success rate of 100% within a confidence range of ^ 0.38. q 2002 Published by Elsevier Science B.V.

AN INTEGRATED SYSTEM FOR HANDWRITTEN DOCUMENT IMAGE PROCESSING

In this paper we attempt to face common problems of handwritten documents such as nonparallel text lines in a page, hill and dale writing, slanted and connected characters. Towards this end an integrated system for document image preprocessing is presented. This system consists of the following modules: skew angle estimation and correction, line and word segmentation, slope and slant correction. The skew angle correction, slope correction and slant removing algorithms are based on a novel method that is a combination of the projection profile technique and the Wigner–Ville distribution. Furthermore, the skew angle correction algorithm can cope with pages whose text line skew angles vary, and handle them by areas. Our system can be used as a preprocessing stage to any handwriting character recognition or segmentation system as well as to any writer identification system. It was tested in a wide variety of handwritten document images of unconstrained English and Modern Greek text from about 100 writers. Add...

Neural classification of abnormal tissue in digital mammography using statistical features of the texture

The authors investigated the efficiency of neural classifiers in recognizing cancer regions of suspicion (ROS) on mammograms. Radial-basis-function (RBF) networks and multilayer perceptron (MLP) neural networks are used to classify ROS including all kinds of abnormalities by processing two types of texture features: statistical descriptors based on high-order statistics and the spatial gray-level dependence (SGLD) matrix. Extensive experiments carried out in the MIAS database have given similar recognition scores for both types of features. The MLP classifier outperforms the score achieved by the RBF networks. Significantly greater training time and computational complexity both in the training and the classification process measured for the MLP networks. Specifically, the recognition accuracy of the MLP is approximately 4% better than that obtained by the RBF networks for the statistical descriptors based on high-order statistics. Using the SGLD matrix the RBF network exceeded the recognition rate of the MLP networks only in one case out of three.