Ioannis Dologlou

Ioannis Dologlou
Institute for Language and Speech Processing | ISLP

Doctor of Engineering

About

78
Publications
2,796
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
551
Citations

Publications

Publications (78)
Conference Paper
This paper presents a system to detect symptoms of allergic rhinitis remotely by using uttered speech and by exploiting its specific spectral characteristics. Based on the principles of adaptive modeling and fundamental frequency variations (jitter) as well as speech analysis by means of acoustic models, the proposed technique achieves an efficient...
Conference Paper
This paper presents a system to detect symptoms of allergic rhinitis remotely by using uttered speech and by exploiting its specific spectral characteristics. Based on the principles of adaptive modelling and fundamental frequency variations (jitter) as well as speech analysis by means of acoustic models, the proposed technique achieves an efficien...
Article
The spectral characteristics of speech can be used to cluster individuals according to whether or notthey suffer from an allergy. Based on the principles of adaptive modelling and fundamental frequencyvariations, as well as speech analysis by means of acoustic models, our technique achieves anefficient classification based on uttered speech over a...
Article
A new algorithm for the design of Hidden Markov Models (HMM) from observed symbol and bisymbol probabilities is presented. The algorithm provides a global optimum and makes use of linear vectorial models of sequences of probabilistic vectors estimated during an off-line learning process. Moreover, a method to enhance observed data so as to comply w...
Article
This work presents our effort to incorporate a state of the art speech recognition engine into a new platform for assistive reading for improving reading ability of Greek dyslexic students. This platform was developed in the framework of the Agent-DYSL, IST project, and facilitates dyslexic children in learning to read fluently. Unlike previously p...
Article
Full-text available
This paper presents a new state-space method for spectral estimation that performs decimation by any factor, it makes use of the full set of data and brings further apart the poles under consideration, while imposing almost no constraints to the size of the Hankel matrix (model order), as decimation increases. It is compared against two previously...
Article
There are several reasons to expect that recognising word order errors in a text will be a difficult problem, and recognition rates reported in the literature are in fact low. Although grammatical rules constructed by computational linguists improve the performance of a grammar checker in word order diagnosis, the repairing task is still very diffi...
Conference Paper
This work presents the evaluation results of a novel technique for word order errors correction, using non native English speakers’ corpus. This technique, which is language independent, repairs word order errors in sentences using the probabilities of most typical trigrams and bigrams extracted from a large text corpus such as the British Nationa...
Conference Paper
In this paper we present an approach for incorporating a state of the art speech recognition engine into a novel assistive reading system for Greek dyslexic students. This system is being developed in the framework of the AGENT-DYSL IST project, and facilitates dyslexic children in learning to read fluently. Unlike previously presented approaches,...
Article
In this paper, two well-known speech enhancement techniques are compared in a Magnetic Resonance Imaging (MRI) scanner noise reduction scheme prior to speech recognition experiment. Our study deals with the comparison between the Non Linear Spectral Subtraction (NSS) with iterative overestimation and the Singular Value Decomposition (SVD)-based noi...
Conference Paper
This paper presents a review of the necessary technology in order to develop a Vocal User interface to be integrated into the jMRUI. jMRUI allows magnetic resonance (MR) spectroscopists to easily perform time-domain analysis of in vivo MR Data and might in the future be used during intraoperative MRI scanning. An operation room with an MRI scanner...
Conference Paper
This paper discusses the improvement of speech recognition in the presence of noise, when a parametric method of signal enhancement is used. The speech enhancement method improves the performance of voice control MRI. This is important since errors in the presence of noise are more frequent and tend to make applications, such as spoken dialogue sys...
Conference Paper
This work presents an evaluation method of Greek sentences with respect to word order errors. The evaluation method is based on words’ reordering and choosing the version that maximizes the number of trigram hits according to a language model. The new parameter of the proposed technique concerns the incorporation of unigram probability. This probab...
Conference Paper
This paper presents an approach for repairing word order errors in English text by reordering words in a sentence and choosing the version that maximizes the number of trigram hits according to a language model. The novelty of this method concerns the use of an efficient confusion matrix technique for reordering the words. For further reducing the...
Conference Paper
What appears to be given in all languages is that words can not be randomly ordered in sentences, but that they must be arranged in certain ways, both globally and locally. The “scrambled” words into a sentence cause a meaningless sentence. Although the use of manually collected grammatical rules can boost the performance of grammar checker in word...
Article
Recognising the verbal content of emotional speech is a difficult problem, and recognition rates reported in the lit- erature are in fact low. Although knowledge in the area has been developing rapidly, it is still limited in fundamental ways. The first issue concerns that not much of the spectrum of emotionally coloured expressions has been studie...
Article
This paper presents an approach for repairing word order errors in English text by reordering words in a sentence and choosing the version that maximizes the number of trigram hits according to a language model. A possible way for reordering the words is to use all the permutations. The problem is that for a sentence with length N words the number...
Article
Full-text available
The objective of this paper is to propose a signal processing scheme that employs subspace-based spectral analysis for the purpose of formant estimation of speech signals. Specifi- cally, the scheme is based on decimative spectral estimation that uses Eigenanalysis and SVD (Singular Value Decompo- sition). The underlying model assumes a decompositi...
Article
This paper deals with the impact of a signal enhancement method on emotional speech recognition confidence scores, in the presence of noise. The emotional speech recognition confidence score reflects the reliability of correctness of the recogniser's output. This is important since errors in the presence of noise are more frequent and tend to make...
Article
This paper presents a new state-space method for spectral estimation that performs decimation by any factor and it is based on Singular Value Decomposition in order to estimate frequency, damping factor, amplitude and phase of complex damped sinusoids in the presence of noise. The new method, called DESE D, makes use of the full set of data and bri...
Article
There are multiple reasons to expect that recognising the verbal content of emotional speech will be a difficult problem, and recognition rates reported in the literature are in fact low. Including information about prosody improves recognition rate for emotions simulated by actors, but its relevance to the freer patterns of spontaneous speech is u...
Article
Full-text available
This paper details on the application of a Decimative Spectral estimation method to speech signals in order to perform spectral analysis and estimation of Formant/Bandwidth values. The method is based on Eigenanalysis and SVD (Singular Value Decomposition) and performs artificial decimation for increased accuracy while it exploits the full set of d...
Article
This paper presents two di#erent directions to build HMM models which give enough acoustic resolution and fit in limited user resources. They both refer to scaling down the acoustic models which are built with tied gaussian HMMs. The total number of gaussians is reduced by a pairwise merging, and the number of gaussians per state is reduced by sele...
Conference Paper
This paper presents a comparison between two parametric methods for Signal Enhancement in order to address the problem of robust Automatic Speech Recognition (ASR). An SVD–based technique (ISE) and a non-linear spectral subtraction method (NSS), have been evaluated by means of the Continuous Speech Recognition system that is used in the ERMIS proje...
Conference Paper
Full-text available
If speech analysis is to detect a speaker's emotional state, it needs to derive information from both linguistic information, i.e., the qualitative targets that the speaker has attained (or approximated), conforming to the rules of language; and paralinguistic information, i.e., allowed variations in the way that qualitative linguistic targets are...
Article
This presentation focuses on the IMUTUS project, which concerns the creation of an innovative method for training users on traditional musical instruments with no MIDI (Musical Instrument Digital Interface) output. The entities collaborating in IMUTUS are ILSP (coordinator), EXODUS, SYSTEMA, DSI, SMF, GRAME, and KTH. The IMUTUS effectiveness is enh...
Article
This paper presents a new state-space method for spectral estimation that performs decimation by any factor D while it imposes no constraints to the model order with respect to D. The new method, called DESED, as well as its Total Least Squares version called DESED_TLS, makes use of the full data set available and is based on SVD in order to estima...
Conference Paper
A new state-space method for spectral estimation that performs decimation by factor two while it makes use of the full set of data available is presented. The proposed method, called DESE2, is based on singular value decomposition in order to estimate frequency, damping factor, amplitude and phase of exponentially damped sinusoids in the presence o...
Conference Paper
This paper presents a new state-space method for spectral estimation based on a companion matrix technique in order to estimate frequency, damping factor, amplitude and phase of exponential sinusoids. The new method, called CSE (companion matrix based spectral estimation), is compared against a previously proposed method called HTLS which is based...
Conference Paper
This paper proposes a new low rate speech coding algorithm, based on a subband approach. At first, a frame of the incoming signal is fed to a low pass filter, thus yielding the low frequency (LF) part. By subtracting the latter from the incoming signal the high frequency (HF), non-smoothed part is obtained. The HF part is modeled using waveform vec...
Article
A scheme for accurate quantification of 1H spectra is presented. The method uses maximum-phase finite impulse response (FIR) filters for solvent suppression and an iterative nonlinear least-squares (NLLS) algorithm for parameter estimation. The estimation algorithm takes the filter influence on the metabolites of interest into account and can there...
Article
This paper presents an iterative signal enhancement algorithm for noise reduction in speech. The algorithm is based on a truncated singular value decomposition (SVD) procedure, which has already been used as a tool for signal enhancement [1][2]. Compared to the classical algorithms, the novel algorithm gives rise to comparable improvements in signa...
Conference Paper
Full-text available
This paper presents two dieren t directions to build HMM models which give enough acoustic resolution and t in limited user resources. They both refer to scaling down the acoustic models which are built with tied gaussian HMMs. The total number of gaussians is reduced by a pairwise merging, and the number of gaussians per state is reduced by select...
Article
This paper proposes a method to transform acoustic models that have been trained with a certain group of speakers for use on different speech in hidden Markov model based (HMM-based) automatic speech recognition. Features are transformed on the basis of assumptions regarding the difference in vocal tract length between the groups of speakers. First...
Article
In this communication we clarify the relationships between the Auto-Regressive models with eXogenous input (ARX) and the dynamic Errors-In-Variables (EIV) models. To this end, we start with the equivalent relationships that can be obtained for the corresponding static models: the Least Squares (LS) model and the Total Least Squares (TLS) model. We...
Article
This paper proposes a new speech coding algorithm, based on a subband approach. In a first step a frame of the incoming signal is fed to a low pass filter. This yields the low frequent (LF) part. By subtracting the latter from the incoming signal, we obtain the high frequent (HF), non-smoothed part. The LF signal is fitted in a Least Squares (LS) s...
Article
We present a new speech coding algorithm, based on an all-pole model of the vocal tract. Whereas current Auto Regressive (AR) based modeling techniques (e.g. CELP, LPC-10) minimize a prediction error, which is considered to be the input to the all-pole model, our approach determines the closest (in L 2 norm) signal, which exactly satisfies an all-p...
Article
In this paper we present a new speech coding algorithm. Like most of the existing vocoders (e.g. CELP, LPC-10, multipulse LPC), it is based on an all-pole model of the vocal tract. The major difference with the existing Linear Predictive Coding (LPC) based algorithms, is the fact that we perform an exact Auto Regressive (AR) modeling. The latter me...
Article
this report is a one-channel algorithm. We think that extending this algorithm to multiple channels (multiple microphones) will improve the performance, because we can exploit the redundancy present in the different channels. In [13] a multiple-channel version of the truncated SVD algorithm has already been presented, so incorporating this version...
Conference Paper
We present a new speech coding algorithm, based on an all-pole model of the vocal tract. Whereas current autoregressive (AR) based modeling techniques (e.g. CELP, LPC-10) minimize a prediction error, which is considered to be the input to the all-pole model, our approach determines the closest (in L<sub>2</sub> norm) signal, which exactly satisfies...
Article
Various signal processing techniques have been proposed to improve spectral estimation of closely spaced sinusoids in the presence of noise. This paper exploits frequency prior knowledge information to extract single peaks in magnetic resonance spectra, corresponding to metabolites of interest, by means of a highly selective finite impulse response...
Article
minimum-phase finite impulse response (FIR) filters for solvent suppression and an iterative nonlinear least-squares (NLLS) algorithm for parameter estimation. The estimation algorithm takes the filter influence on the metabolites of interest into account and can thereby correctly incorporate a large variety of prior knowledge in the estimation pha...
Article
This paper describes the permutative vector quantization (PVQ) scheme as a special case of a more general structurally constrained vector quantization concept. This concept makes it possible to increase the vector dimensions beyond the technical bounds of conventional VQ and to exploit by these means, the inter-pixel correlations in large image blo...
Article
Various signal enhancement techniques have been proposed to improve spectral estimation of closely spaced sinusoids in the presence of noise. This correspondence exploits the filtering interpretation of some well-known SVD-based iterative algorithms for signal enhancement to further improve their performance. It is shown that the initial conditions...
Article
In this paper, it is shown how the advantages of continuous regularization (CR) can be exploited to achieve an improved, fully automated LPSVD analysis of MRS time-domain data. The main advantage of CR is its ability to determine the number of spectral components even at low signal-to-noise ratios, which suggested its use forin vivospectroscopy. Es...
Article
This work concerns estimation of optimal MRI images from incomplete raw data. 'Raw' implies that the data have not yet been Fourier transformed to the spatial (image) domain, whereas 'incomplete' pertains to the fact that a number of points on the Cartesian sampling grid were omitted in order to reduce the scan time. We introduce a novel SVD-based...
Article
This paper presents a new decomposition for exact representation of a signal in terms of sinusoids with arbitrary frequencies. It is shown that the choice of N2 distinct frequencies can provide a set of N linearly independent sine and cosine vectors in the N-dimensional space which in turn enable the exact decomposition of any N samples signal. A d...
Article
This paper proposes a methodology for designing linear models for multichannel signals. Despite the interest of conventional approaches, these methods are not always well adapted to spectral estimation and enhancement via global rank reduction techniques. It is shown that these problems may however be solved by the new design algorithm proposed her...
Conference Paper
We show how successive projection-like algorithms may be used for approximation or exact modelling of a signal. For this purpose, we propose a new efficient algorithm providing adequate linear difference equations satisfied by the original signal. The projection operators at each step of the approximation algorithm and the new procedure are shown t...
Conference Paper
The paper describes the permutative vector quantization (PVQ) scheme as a special case of a more general structurally constrained vector quantization concept. This concept makes it possible to increase the vector dimensions beyond the technical bounds of conventional VQ and to exploit, by means of this, the inter-pixel correlations in large image b...
Article
A zero error modeling technique for two-dimensional signals is presented. It is pointed out that a rank reduction algorithm applied to the signal renders the modeling physically meaningful in the sense that the estimated exponential sinusoides coincide with the real ones. Image compression is performed as an application of the present method.
Article
Une méthode de modélisation d'images par sinusoïdes exponentielles est présentée. La séparabilité du problème est soulignée par la décomposition en valeurs singulières de l'image ce qui entraîne que les paramètres peuvent être déterminés à partir de deux signaux multicanal. Un schéma de compression d'images est élaboré comme une application possibl...
Article
This paper presents a new matrix formulation of the basic concepts governing discrete Hidden Markov Models (HMM). Using this formulation, we show that symbol and state probabilities are exponential functions of the transition matrix of the model. Furthermore, based on the eigenanalysis of the transition matrix. a closed form relationship is derived...
Article
An enhanced version of a signal is obtained after subtraction of a linear combination of rank one matrices from the signal observation matrix. This operation results in a nonToeplitz observation matrix. By averaging the elements of this matrix, a new Toeplitz matrix is produced and a reconstructed signal is generated. This averaging operation is ex...
Article
The frequency analysis of a signal is reconsidered in this paper. Although the classical approach based on the representation of a signal in terms of complex exponentials has its merit, an alternative approach aiming to represent a signal by a sum of exponential functions is also interesting. An exact solution to this problem can be found at the me...
Conference Paper
The frequency analysis of a signal is considered. Although the classical approach based on the representation of a signal in terms of complex exponentials has its merit, an alternative approach aiming to represent a signal by a sum of exponential functions is also interesting. An exact solution to this problem is found at the meeting point of linea...
Article
Full-text available
A new idea for waveform coding using vector quantisation (VQ) is introduced. This idea makes it possible to deal with codevectors much larger than before for a fixed bit per sample rate. Also a solution to the matching problem (inherent in the present context) in the L<sub> infinity </sub>-norm describing a measure of nearness is presented. The ove...
Conference Paper
Full-text available
This paper presents an experimental preference tool designed, implemented and tested in the Eurotra project. The mechanism is based on preference rules which can either compare subtrees pairwise or single out a subtree on the basis of some specified constraints. Scoring permits combining the effects of various preference rules.
Article
A new algorithm for the extraction of the fundamental frequency time-varying information is described. The algorithm is based on the iterative use of a linear filter with zero phase and monotonically decreasing frequency response (low pass). The results show that the method is both efficient and robust in noisy environments, providing an estimate f...
Conference Paper
Full-text available
A new idea for waveform coding using vector quantization (VQ) is introduced. This idea makes it possible to deal with codevectors much larger than before for a fixed bit per sample rate. Also a solution to the matching problem in the L<sub>∞</sub>-norm describing a measure of nearness is presented. The overall computational complexity of the soluti...
Conference Paper
Full-text available
A new idea for waveform coding using vector quantization (VQ) is introduced. This idea makes it possible to deal with codevectors much larger than before for a fixed bit per sample rate. Also a solution to the matching problem in the L∞-norm describing a measure of nearness is presented. The overall computational complexity of the solution is O(n3l...
Article
There are multiple reasons to expect that detecting the word order errors in a text will be a difficult problem, and detection rates reported in the literature are in fact low. Although grammatical rules constructed by computer linguists improve the performance of grammar checker in word order diagnosis, the repairing task is still very difficult....
Article
A good indicator of whether a person really knows the context of language is the ability to use in correct order the appropriate words in a sentence. The "scrambled" words cause a meaningless and ill formed sentences. Since the language model, is extracted from a large text corpus, it encodes the local dependencies of words. The word order errors u...
Article
This paper deals with the impact of a signal enhancement method on speech recognition confidence scores, in the presence of noise. The speech recognition confidence score reflects the reliability of correctness of the recogniser's output. This is important since errors in the presence of noise are more frequent and tend to make applications, such a...
Article
Speech recognition errors affect the performance of multimodal systems especially in noisy conditions. For that reason the use of speech recognition confidence score and time stamps is suggested. This paper deals with the impact of a signal enhancement method on these speech recognition's parameters, in the presence of noise. By recognising noisy s...
Article
Full-text available
We propose a method to transform the on line speech signal so as to comply with the specica-tions of an HMM-based automatic speech recog-nizer. The spectrum of the input signal undergoes a v ocal tract length (VTL) normalization based on dierences of the average third formant F 3 . The high frequency gap which is generated after scaling is estimate...
Article
Nous présentons une nouvelle formulation du problème de la modélisation multicanale qui repose sur une description compacte, au moyen d'une représentation d'état. L'optimisation des paramètres du modèle porte sur ses propriétés fréquentielles. Par ailleurs, nous proposons un algorithme de projections successives qui permet d'améliorer les caractéri...

Network

Cited By