| 摘要 |
<p>In a 3 dimension model-based coding system, a shape of mouth, especially, a shape of lip, among a shape of face, has an important visual information. As each person has the different shape of lip, however, it is desirable to transform the shape of lip based on an individual person. First, a characterized 3 dimension lip model based on the individual person is generated 55 by matching a 3 dimension lip model 50 to a basic lip 45 of the individual person. Subsequently, the characterized 3 dimension lip model is converted 60 into a 2 dimension lip model and rotated according to a angle of a real lip image L20. In case that the person is talking, the 2 dimension lip model is reconstructed 35 based on a recognized syllable 30. Finally, a difference between the 2 dimension lip model and the real lip image is calculated. As not a whole real lip image but the difference between the real image and the 2 dimension lip model similar to the real lip image is transmitted, resultingly, the amount of transmission data can be reduced. Specifically lip movement parameters 70, voice signals L10, and lip model rotation angles 60 are transmitted from MUX 75.</p> |
