Analysis and Synthesis of Facial Expressions - 3D Modeling and Animation: Synthesis and Analysis Techniques for the Human Body - page 252

analysis and synthesis. Experimental results from the approach in Eisert (2000)
are provided in order to illustrate the applicability of model-based techniques to
these applications.
Model-Based Coding
In recent years, several video coding standards, such as H.261/3 and MPEG-1/
2/4 have been introduced to address the compression of digital video for storage
and communication services. These standards describe a hybrid video coding
scheme, which consists of block-based motion-compensated prediction (MCP)
and DCT-based quantification of the prediction error. The recently determined
H.264 standard also follows the same video coding approach. These waveform-
based schemes utilize the statistics of the video signal without knowledge of the
semantic content of the frames and achieve compression ratios of several
hundreds-to-one at a reasonable quality.
If semantic information about a scene is suitably incorporated, higher coding
efficiency can be achieved by employing more sophisticated source models.
Model-based video codecs, e.g., use 3-D models for representing the scene
content. Figure 5 shows the structure of a model-based codec for the application
of video telephony.
Figure 5: Structure of a model-based codec.