Summary: | 碩士 === 國立臺灣大學 === 資訊工程學系 === 86 === The proposed system is named Image Talk: a real-time synthetic
talking head using one single image with Chinese text-to-speech
capability. Image Talk uses one single image to automatically
create video-like talking sequences in real time. The image
can be acquired from a photograph, video clip, or hand drawn
characters. This interactive system accepts Chinese text input
and talks back in Mandarin Chinese, generating facial
expression in real time.
Image Talk analyzes Chinese text by converting it to a standard
Pinyin system used in Taiwan and fetches the associated facial
expressions from an expression pool dynamically. The
expressions are synchronized with the synthetic speech and
played back in the video-like talking sequence in real time.
Image Talk also incorporates eye blinking, small-scale head
rotation and translation perturbations, to make the results
more natural. It is also easy to switch to any other facial or
non-facial images, such as animals or machines for special
effects. The result is quite entertaining, and can easily be
used as a new human-machine interface, as well as for lip
synchronization for computer animated characters.
|