Loading...
42 results
Search Results
Now showing 1 - 10 of 42
- Evaluation of a segmental durations model for TTSPublication . Teixeira, João Paulo; Freitas, Diamantino SilvaIn this paper we present a condensed description of a European Portuguese segmental duration’s model for TTS purposes and concentrate on its evaluation. This model is based on artificial neural networks. The evaluation of the model quality was made by comparison with read speech. The standard deviation reached in test set is 19.5 ms and the linear correlation coefficient is 0.84. The model is perceptually evaluated with 4.12 against 4.30 for natural human read speech in a scale of 5.
- Electroencephalogram cepstral distances in alzheimer’s disease diagnosisPublication . Rodrigues, Pedro Miguel; Freitas, Diamantino Silva; Teixeira, João PauloAlzheimer's disease (AD) represents one ofthe greatest public health challenges worldwide nowadays, because it affects millions of people ali o ver the world and it is expected that the disease will increase considerably in the near future. This study is the first application attempt of cepstral analysis on Electroencephalogram (EEG) signals to find new parameters in arder to achieve a better differentiation belween EEGs of AD patients and Control subjects. The results show that the methodology that uses a combined Wavelet (WT) Biorthogonal (Bior) 3.5 and cepstrum analysis was able to describe the EEG dynamics with a higher discriminative power than the other WTs/spectmm methodologies m previous studies. The most important significance figures were found in cepstral distances between cepstrums oftheta and alpha bands (p=0. 00006<0. 05).
- Back close non-syllabic vowel [U] behaviour in European Portuguese: reduction or suppression?Publication . Barros, Maria João; Braga, Daniela; Freitas, Diamantino Silva; Teixeira, João Paulo; Latsch, VagnerIn this paper we analyse and describe the phonetic behavior of the non-tonic back closed vowel [u] in European Portuguese (EP) in what respects its quality and relative duration. Slow and faster speech was used. The objective of this work is to contribute with experimental results in order to obtain a more accurate comprehension of this vowel’s profile, which can be integrated and extended to other non-tonic vowels’ behaviors in the EP non-tonic vowel system. This may represent an important improvement in synthetic speech quality and naturalness, as it concerns acoustical parameters, rhythm and supra-segmental features, which means prosody. This study is inscribed in a project in Speech Synthesis for EP held by an inter-disciplinary group in intimate articulation between the engineering experience and tools and the linguistic approach.
- Phoneme dedicated ANN improves segmental duration modelPublication . Teixeira, João Paulo; Freitas, Diamantino Silva
- Early detection of electroencephalogram temporal events in Alzheimer's diseasePublication . Rodrigues, Pedro Miguel; Freitas, Diamantino Silva; Teixeira, João Paulo; Alves, Dílio; Garrett, CarolinaAlzheimer’s Disease (AD) is considered one of the most debilitating illness in modern societies and the leading cause of dementia. This study is a new approach to detect early AD Electroencephalogram (EEG) temporal events in order to improve early AD diagnosis. For that, Self-Organized Maps (SOM) were used, and it was found that there are sequences of EEG energy variation, characteristic of AD, that appear with high incidence in Mild Cognitive Impairment (MCI) patients. Those AD events are related to the first cognitive changes in patients that interfered with the normal EEG signal pattern. Moreover, there are significant differences concerning the propagation time of those events between the study groups(p=0.0082<0.05), meaning that, as AD progresses the brain dynamics are progressively affected, what is expected because AD causes brain atrophy.
- PRO-GMÁTICA: uma base de dados prosódica pragmaticamente orientada em Português EuropeuPublication . Braga, Daniela; Teixeira, João Paulo; Coelho, Luís; Freitas, Diamantino SilvaNeste trabalho apresenta-se uma base de dados de fala espontânea com etiquetagem linguística, prosódica e pragmática em Português Europeu (PE). Esta informação pode ser encontrada num ficheiro de texto que é associado ao sinal de fala, sendo por isso possível analisar simultaneamente as várias dimensões em questão. O objectivo deste trabalho consiste em criar um corpus de fala etiquetado em que seja possível correlacionar a dimensão acústica e prosódica do sinal com o seu significado linguístico e pragmático, sobretudo tendo em atenção as dimensões emocionais e estilísticas do discurso. Constituem motivações subjacentes a este trabalho a necessidade de controlar a manipular a naturalidade da fala sintética, a necessidade de implementar emoção e atitude nos actuais sistemas síntese e reconhecimento de fala e a inexistência de corpora de fala espontânea linguística e prosodicamente etiquetada e que contemple as dimensões emocional e comunicativa. Neste artigo, faz-se uma breve introdução ao assunto em questão e apresenta-se o estado-da-arte das bases de dados em PE. Em seguida, descreve-se o processo da recolha da PRO-GMÁTICA e os critérios subjacentes à sua estruturação. Posteriormente, refere-se o tratamento e análise dos dados recolhidos, o que consiste basicamente no elenco e descrição dos níveis de etiquetagem considerados e na extracção dos parâmetros acústicos segundo o modelo Fujisaki. Finalmente, apresentam-se as conclusões e trabalhos futuros. Desta forma, pretendemos, num futuro próximo, vir a dotar os sistemas de síntese da fala e de diálogo homem-máquina em PE de uma dimensão mais pragmática.
- Accoustical characterisation of accented syllable in portuguese; a contribution to the naturalness of speech synthesisPublication . Teixeira, João Paulo; Paulo, Elisabete R.; Freitas, Diamantino Silva; Pinto, Maria da GraçaText-to-Speech systems require control of the prosodic parameters of the produced speech waveform in order to achieve a higher naturalness and degree of perception. Amongst the several dimensions into which prosody can be unfolded, the accented syllable realisation brings the basic problem of producing a set of comprehensive rules for accurate control of the acoustic realisation of the syllable parameters, which remains to be solved. In particular, for the Portuguese language, in Europe, a set of comprehensive quantitative characterisation data and rules is absolutely lacking. The present paper is intended, as a quantitative contribution, as far as we know the first, to the solution of this problem. The duration Intensity, and variation of F0 were modelled in the tonic syllable according to its position in the word and the position of the word in the sentence.
- Editorial: advances in machine learning approaches and technologies for supporting nervous system disease diagnosisPublication . Rodrigues, Pedro Miguel; Bispo, Bruno; Freitas, Diamantino Silva; Marques, João Alexandre Lobo; Teixeira, João PauloThe nervous system is essential for physical and mental health but is complex and delicate. As it can unfortunately be affected by several progressive diseases, an early diagnosis is often critical for effective treatment (Xu et al., 2022). The diagnosis of nervous system diseases traditionally relies on a combination of clinical examination, imaging and signals tests, and laboratory tests (Siuly and Zhang, 2016). However, these methods can be time-consuming, expensive, and not always accurate (Milligan, 2019). In an era marked by unprecedented technological advances in machine learning (ML), a computational tool that allows the identification of patterns in data that would be difficult or even impossible for humans, its application to assist in medical diagnosis emerges as a beacon of hope in the complex panorama of nervous system diseases. The Research Topic Advances in machine learning approaches and technologies for supporting nervous system disease diagnosis aims to shed light on the transformative role that ML-based approaches and technologies are playing in reshaping the way an ensemble of nervous system disorders are understood, diagnosed, and treated.
- Alzheimer’s early prediction with electroencephalogramPublication . Rodrigues, Pedro Miguel; Teixeira, João Paulo; Garrett, Carolina; Alves, Dílio; Freitas, Diamantino SilvaAlzheimer's disease (AD) is currently an incurable illness that causes dementia and patienfs condition is progressively worse and it represents one ofthe greatest public health challenges worldwide. The main objective ofthis work was to develop a classificatiwmethodology for EEG signals to improve discrimination amongst patients at varying stages ofthe illness, Mitd Cognitive Impairment (MCI) patients and non-patients either in order to obtain a more reliable methodology to identify AD in early stages.
- Prediction of Fujisaki model’s phrase commandsPublication . Teixeira, João Paulo; Freitas, Diamantino Silva; Fujisaki, HiroyaThis paper presents a model to predict the phrase commands of the Fujisaki Model for F0 contour for the Portuguese Language. Phrase commands location in text is governed by a set of weighted rules. The amplitude (Ap) and timing (T0) of the phrase commands are predicted in separate neural networks. The features for both neural networks are discussed. Finally a comparison between target and predicted values is presented.