Appeal No. 1997-3225
Application No. 08/351,218
through 72 and 79 through 84. Claims 8, 10, 38 through 40,
53, 55 and 73 through 78 were canceled.
The invention relates to a technique for allowing a
computer to simulate an animated image of a human speaking.
As identified in the specification on pages 9 and 10, a spoken
language is broken down into units of speech, phonemes and
diphthongs. Then as identified on page 10 of the
specification, a video of a person speaking is obtained and
individual frames which best represent the person using these
units of speech are determined. As identified on pages 11
through 14 of the specification, these frames are then saved
in a database, each image corresponding to a phoneme. As
described on pages 14 and 15 of the specification, features in
each image in the database are matched to identical features
in the other images. These features in the images are called
tiepoints. As identified on page 15 of the specification, an
audio speech sequence to be synchronized with the images is
then analyzed to determine spoken phonemes and their relative
timing. Then as identified on pages 15 through 18 and 32
2
Page: Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Next
Last modified: November 3, 2007