| Abstract [eng] |
Speech animation is widely used in technical devices to allow the growing number of hearing impaired persons, children, middle-aged and elderly equal participation in communication. Accuracy of the speech recognition or generation engine, modelled visemes, phonemes to visemes mapping and coarticulation control model mainly influence speech animation quality, so phonetics of specific language must be analysed to create new „Talking head” (synthetic head model that can produce visual speech). Framework suitable to visualize Lithuanian language, which includes two new models to improve intelligibility of animated speech is proposed and applied to create Lithuanian „Talking head” „LIT”. The dissertation consists of Introduction, three main chapters and general conclusions. Chapter 1 provides the analysis of the existing speech animation technologies. Different facial modelling techniques are examined to define the most suitable 3D „Talking head” modelling technique. Viseme classification experiments across different languages are analysed to identify variety of viseme identification and classification methods. Different coarticulation control models are compared to decide which one should be used to define coarticulation of Lithuanian speech. Chapter 2 describes theoretical framework for Lithuanian speech animation. Translingual visual speech synthesis methods, specifics of Lithuanian phonetics and phonemes recognition system which is included in the proposed framework are analysed. Chapter 3 presents the proposed Lithuanian speech animation technology and two new models to improve intelligibility of animated Lithuanian speech. The new translingual viseme selection model divides viseme classes of any language into two parts: constant (translingual) and variable. Technology to create constant (translingual) viseme set for 22 most popular consonants is proposed. Coarticulation control model specified for pure Lithuanian diphthongs is also proposed. The chapter is concluded with practical evaluation of the Lithuanian „Talking head” „LIT”. General conclusions summarise the present study. It is followed by an extensive list of 144 references and a list of 3 publications by the author on the topic of the dissertation. |
