Kyriakos N. Sgarbas

Handwritten word recognition based on structural characteristics and lexical support

In this paper a handwritten recognition algorithm based on structural characteristics, histograms and profiles, is presented. The well-known horizontal and vertical histograms are used, in combination with the newly introduced radial histogram, out-in radial and in-out radial profiles for representing 32 /spl times/ 32 matrices of characters, as 280-dimension vectors. The recognition process has been supported by a lexical component based on dynamic acyclic FSAs (Finite-State-Automata).

TWO ALGORITHMS FOR INCREMENTAL CONSTRUCTION OF DIRECTED ACYCLIC WORD GRAPHS

In this paper we present two algorithms for building lexicons in Directed Acyclic Word-Graphs (DAWGs). The two algorithms, one for deterministic and the other for non-deterministic DAWGs, can be used instead of the traditional subset construction method. Although the proposed algorithms do not produce the optimal DAWG (i.e., the one with the minimum number of states), they are simple, fast and able to build the DAWG incrementally, as new words are added to the lexicon. Thus, building large lexicons in a DAWG structure becomes an easy task, even for a modest computer.

Optimal insertion in deterministic DAWGs

In this paper, we present an on-line algorithm for adding words (strings) in deterministic directed acyclic word graphs (DAWGs) i.e. acyclic deterministic finite-state automata (DFAs). The proposed algorithm performs optimal insertion, meaning that if applied to a minimal DAWG, the DAWG after the insertion will also be minimal. The time required to add a new word is O(n) with respect to the size of the DAWG. Repetitive application of the proposed insertion algorithm can be used to construct minimal deterministic DAWGs incrementally, although the algorithm is not time-efficient for building minimal DAWGs from a set of words: to build a DAWG of n words this way, O(n2) time is required. However, the algorithm is quite useful in cases where existing minimal DAWGs have to be updated rapidly (e.g. speller dictionaries), since each word insertion traverses only a limited portion of the graph and no additional minimization operation is required. This makes the process very efficient to be used on-line. This paper provides a proof of correctness for the algorithm, a calculation of its time-complexity and experimental results.

Speaker Identification Using Semi-supervised Learning

Semi-supervised classification methods use available unlabeled data, along with a small set of labeled examples, to increase the classification accuracy in comparison with training a supervised method using only the labeled data. In this work, a new semi-supervised method for speaker identification is presented. We present a comparison with other well-known semi-supervised and supervised classification methods on benchmark datasets and verify that the presented technique exhibits better accuracy in most cases.

Dynamic gene network reconstruction from gene expression data in mice after influenza A (H1N1) infection.

BackgroundThe immune response to viral infection is a temporal process, represented by a dynamic and complex network of gene and protein interactions. Here, we present a reverse engineering strategy aimed at capturing the temporal evolution of the underlying Gene Regulatory Networks (GRN). The proposed approach will be an enabling step towards comprehending the dynamic behavior of gene regulation circuitry and mapping the network structure transitions in response to pathogen stimuli.ResultsWe applied the Time Varying Dynamic Bayesian Network (TV-DBN) method for reconstructing the gene regulatory interactions based on time series gene expression data for the mouse C57BL/6J inbred strain after infection with influenza A H1N1 (PR8) virus. Initially, 3500 differentially expressed genes were clustered with the use of k-means algorithm. Next, the successive in time GRNs were built over the expression profiles of cluster centroids. Finally, the identified GRNs were examined with several topological metrics and available protein-protein and protein-DNA interaction data, transcription factor and KEGG pathway data.ConclusionsOur results elucidate the potential of TV-DBN approach in providing valuable insights into the temporal rewiring of the lung transcriptome in response to H1N1 virus.

A low cost approach for brain tumor segmentation based on intensity modeling and 3D Random Walker

Abstract Objective Magnetic resonance imaging (MRI) is the primary imaging technique for evaluation of the brain tumor progression before and after radiotherapy or surgery. The purpose of the current study is to exploit conventional MR modalities in order to identify and segment brain images with neoplasms. Methods Four conventional MR sequences, namely, T1-weighted, gadolinium-enhanced T1-weighted, T2-weighted and fluid attenuation inversion recovery, are combined with machine learning techniques to extract global and local information of brain tissues and model the healthy and neoplastic imaging profiles. Healthy tissue clustering, outlier detection and geometric and spatial constraints are applied to perform a first segmentation which is further improved by a modified multiparametric Random Walker segmentation method. The proposed framework is applied on clinical data from 57 brain tumor patients (acquired by different scanners and acquisition parameters) and on 25 synthetic MR images with tumors. Assessment is performed against expert-defined tissue masks and is based on sensitivity analysis and Dice coefficient. Results The results demonstrate that the proposed multiparametric framework differentiates neoplastic tissues with accuracy similar to most current approaches while it achieves lower computational cost and higher degree of automation. Conclusion This study might provide a decision-support tool for neoplastic tissue segmentation, which can assist in treatment planning for tumor resection or focused radiotherapy.

APPLYING SIMILARITY MEASURES FOR AUTOMATIC LEMMATIZATION: A CASE STUDY FOR MODERN GREEK AND ENGLISH

This paper addresses the problem of automatic induction of the normalized form (lemma) of regular and mildly irregular words with no direct supervision using language-independent algorithms. More specifically, two string distance metric models (i.e. the Levenshtein Edit Distance algorithm and the Dice Coefficient similarity measure) were employed in order to deal with the automatic word lemmatization task by combining two alignment models based on the string similarity and the most frequent inflectional suffixes. The performance of the proposed model has been evaluated quantitatively and qualitatively. Experiments were performed for the Modern Greek and English languages and the results, which are set within the state-of-the-art, have showed that the proposed model is robust (for a variety of languages) and computationally efficient. The proposed model may be useful as a pre-processing tool to various language engineering and text mining applications such as spell-checkers, electronic dictionaries, morphological analyzers etc.

Incremental Construction of Compact Acyclic NFAs

This paper presents and analyzes an incremental algorithm for the construction of Acyclic Non-deterministic Finite-state Automata (NFA). Automata of this type are quite useful in computational linguistics, especially for storing lexicons. The proposed algorithm produces compact NFAs, i.e. NFAs that do not contain equivalent states. Unlike Deterministic Finite-state Automata (DFA), this property is not sufficient to ensure minimality, but still the resulting NFAs are considerably smaller than the minimal DFAs for the same languages.

Joint Spatial-Spectral Feature Space Clustering for Speech Activity Detection from ECoG Signals

Brain-machine interfaces for speech restoration have been extensively studied for more than two decades. The success of such a system will depend in part on selecting the best brain recording sites and signal features corresponding to speech production. The purpose of this study was to detect speech activity automatically from electrocorticographic signals based on joint spatial-frequency clustering of the ECoG feature space. For this study, the ECoG signals were recorded while a subject performed two different syllable repetition tasks. We found that the optimal frequency resolution to detect speech activity from ECoG signals was 8 Hz, achieving 98.8% accuracy by employing support vector machines as a classifier. We also defined the cortical areas that held the most information about the discrimination of speech and nonspeech time intervals. Additionally, the results shed light on the distinct cortical areas associated with the two syllables repetition tasks and may contribute to the development of portable ECoG-based communication.

Using the Levenshtein Edit Distance for Automatic Lemmatization: A Case Study for Modern Greek and English

In the present work we have implemented the Edit Distance (also known as Levenshtein Distance) on a dictionary-based algorithm in order to achieve the automatic induction of the normalized form (lemma) of regular and mildly irregular words with no direct supervision. The algorithm combines two alignment models based on the string similarity and the most frequent inflexional suffixes. In our experiments, we have also examined the language-independency (i.e. independency of the specific grammar and inflexional rules of the language) of the presented algorithm by evaluating its performance on the Modern Greek and English languages. The results were very promising as we achieved more than 95 % of accuracy for the Greek language and more than 96 % for the English language. This algorithm may be useful to various text mining and linguistic applications such as spell-checkers, electronic dictionaries, morphological analyzers, search engines etc.