Generating animation from visual and audio input
A technique for generating an animated character based on visual and audio input from a live subject. Further described is a technique of extracting phonemes to select corresponding visemes to model a set of physical positions of the subject or emotional expression of the subject.
1. A machine-readable medium having stored thereon a set of instructions, which when executed by a machine cause the machine to perform a method comprising;
receiving a visual input and a speech input;
extracting a phoneme from the speech input; and
identifying a plurality of emotional visemes based upon the phoneme;
transforming one of the plurality of emotional visemes to generate a transformed emotional viseme.
2. The machine-readable medium of claim 1 further comprising computing an optical flow of the visual input.
3. The machine-readable medium of claim 2 further comprising classifying the optical flow according to an expression type.
4. The machine-readable medium of claim 3 further comprising selecting one of the plurality of emotional visemes corresponding to the expression type.
5. The machine-readable medium of claim 1 further comprising estimating an affine model according the visual input.
6. The machine-readable medium of claim 5 further comprising morphing the transformed emotional viseme with other emotional visemes.
7. The machine-readable medium of claim 1 further comprising selecting a neutral viseme corresponding to the phoneme.
8. The machine-readable medium of claim 7 wherein the neutral viseme is used to identify the plurality of emotional visemes.
9. A method comprising;
extracting a plurality of phonemes from a speech input;
selecting a plurality of visemes based upon the plurality of phonemes;
transforming the plurality of visemes according to a visual input to generate a plurality of transformed visemes; and
morphing the plurality of transformed visemes,
wherein the visemes are emotional visemes to be modified according to emotional expression types.
10. The method of claim 9 wherein the plurality of visemes are neutral visemes.
11. The method of claim 10 wherein the neutral visemes are transformed according to a plurality of affine models.
12. The method of claim 9 wherein the plurality of visemes are emotional visemes.
13. The method of claim 12 wherein the emotional visemes are transformed according to a plurality of expression types.
14. An apparatus comprising:
a first unit to extract phonemes from a speech input;
a second unit to select visemes corresponding to the phonemes; and
a third unit to modify the visemes according to a visual inputs,
wherein the visemes are emotional visemes to be modified according to emotional expression types.
15. The apparatus of claim 14 wherein the visemes are neutral visemes to be modified according to an affine model.
16. The apparatus of claim 15 wherein the affine models correspond to regions of the visual input.
17. The apparatus of claim 16 wherein the neutral visemes are to be morphed with other visemes after the neutral visemes are modified.
18. The apparatus of claim 14 wherein the emotional expression types correspond to an optical flow of the visual input.
19. The apparatus of claim 18 wherein the emotional visemes are to be morphed with other visemes.
20. A system comprising:
an image capturing device to capture a plurality of frames of a visual input;
a sound capturing device to capture a plurality of samples of a sonic input;
an extraction device to extract phonemes from the plurality of samples of the sonic input;
a selection device to select visemes corresponding to the phonemes;
a tracking device to track regions of the visual input;
a transforming device to spatially transform the visemes according to an affine model; and
a morphing device to morph the visemes after they are transformed by the transforming device.
21. The system of claim 20 wherein the visemes are neutral visemes.
22. The system of claim 20 wherein the visemes are emotional visemes to be selected according to a classification of expression type of the plurality of frames of visual input.
23. The system of claim 22 further comprising an optical flow computing device to compute an optical flow of the plurality of frames of the visual input.