“Variable rate speech animation synthesis” by Kubo, Adachi, Terzopoulos and Morishima

  • ©Akane Yano, Hiroyuki Kubo, Yoshihiro Adachi, Demetri Terzopoulos, and Shigeo Morishima

  • ©Akane Yano, Hiroyuki Kubo, Yoshihiro Adachi, Demetri Terzopoulos, and Shigeo Morishima

Conference:


Type(s):


Title:

    Variable rate speech animation synthesis

Presenter(s)/Author(s):



Abstract:


    Speech animation has traditionally been achieved by two main approaches: image-based methods [Ezzat et al. 2002] and key framing methods [Cohen and Massaro 1993]. Both approaches require large databases that include many images or 3D shapes with facial expressions and speech lip shapes. Recently, mocap-based facial animation techniques have been utilized to create natural speech animation. However, it is difficult to create natural lip postures with variable speech rates by reusing the motions of pre-captured lip markers. To ameliorate these problems, we present a novel lip-sync interpolation technique that supports flexible speech rates and does not need large databases. Our system can synthesize speech animation with variable speech posture from input speech.


Additional Images:

©Akane Yano, Hiroyuki Kubo, Yoshihiro Adachi, Demetri Terzopoulos, and Shigeo Morishima

ACM Digital Library Publication:



Overview Page: