Search

In this paper we present a silent speech interface (SSI) system aimed at restoring speech communication for individuals who have lost their voice due to laryngectomy or diseases affecting the vocal folds. In the proposed system, articulatory data captured from the lips and tongue using permanent magnet articulography (PMA) are converted into audible speech using a speaker-dependent transformation learned from simultaneous recordings of PMA and audio signals acquired before laryngectomy. The transformation is represented using a mixture of factor analysers, which is a generative model that allows us to efficiently model non-linear behaviour and perform dimensionality reduction at the same time. The learned transformation is then deployed during normal usage of the SSI to restore the acoustic speech signal associated with the captured PMA data. The proposed system is evaluated using objective quality measures and listening tests on two databases containing PMA and audio recordings for normal speakers. Results show that it is possible to reconstruct speech from articulator movements captured by an unobtrusive technique without an intermediate recognition step. The SSI is capable of producing speech of sufficient intelligibility and naturalness that the speaker is clearly identifiable, but problems remain in scaling up the process to function consistently for phonetically rich vocabularies. [ABSTRACT FROM AUTHOR]

Abstract: Modelling human tutors' socio-linguistic abilities computationally is necessary to underpin human–computer educational interactions with best pedagogic practices and to enhance the naturalness of such interactions. In this paper we present a computational model of tutorial feedback selection based on the context of the immediate situation for which the feedback is selected as well as on politeness considerations shown to be of importance to increasing pedagogical efficacy of computer assisted learning (Wang et al., 2008). The model is based on empirical studies with human tutors and on extensive linguistic analysis of human tutorial dialogues collected with the specific aim to inform the implementation of a natural language tutorial dialogue interface. The evaluation of the model, involving the comparison of its output with the linguistic responses produced by a human tutor, demonstrates the model's plausibility and highlights future directions for improving natural language human–computer interactions for educational purposes. [Copyright &y& Elsevier]