| Summary: | 碩士 === 國立清華大學 === 資訊工程學系 === 86 === Recently, the popularity of Internet has promoted the
development of Internet-based applications.People who use these
applications always need to have a virtual representative such
as "user ID" or "user icon" etc..Due to the advance of software
and hardware technology, the use of 3D virtual human as an
avatar in the application has been studied extensively. While
applications use virtual humans as virtual representatives,
people expect both the movements of bodies and the facial
expressions. The lip movement during speaking plays a very
important role incommunication. There are many researches in
studying the synchronization between lip shape and speech.
However, these researches mostly focus on English.In this
thesis, we proposed a software system architecture for
generating mandarin lip animation. Besides, we implement a
system called M-Lips (Mandarin Lip Animation System) based on
our architecture. First, a Chinese viseme table is analyzed
according to the Chinese phonetic characteristics and it has
been classified into 12 standard Chinese visemes.Then an input
text is decomposed into syllables for retrieving the
corresponding viseme numbers in the "syllable to viseme" mapping
table. Moreover, we use cubic Bezier curve to model the lip and
then calculate the nodal trajectory of lip by using dynamic
nodal displacement algorithm. Finally, we integrated M-Lips with
the Chinese speech synthesizer.The experiment system
satisfactorily reads the given text and converts it to speech
with lip animation.
|
|---|
