Realistic Korean speech modeling and synchronization

Summary form only given. Speech synthesis plays an important role in Korean speech animation since it produces the sequence of phonetic phonemes by processing nearly all of the elements like phonological rules, to enunciate the input text. To make a speaking facial model which uses the motions of lips and surrounding part of the face, it is necessary to extract visual informations from the output of a speech synthesis system. The paper investigates a set of facial elements enough to make a realistic speech animation by examining the visual factors not available from speech synthesis systems as well as extracting signals from the result of speech synthesis that will cause visible changes in facial expression.