Systems and methods for animating non-humanoid characters with human motion data
View Patent ↗Systems, methods and products for animating non-humanoid characters with human motion are described. One aspect includes selecting key poses included in initial motion data at a computing system; obtaining non-humanoid character key poses which provide a one to one correspondence to selected key poses in said initial motion data; and statically mapping poses of said initial motion data to non-humanoid character poses using a model built based on said one to one correspondence from said key poses of said initial motion data to said non-humanoid character key poses. Other embodiments are described.
1. A method for animation comprising:
selecting key poses included in initial motion data at a computing system;
obtaining non-humanoid character key poses which provide a one to one correspondence to selected key poses in said initial motion data; and
statically mapping poses of said initial motion data in a first observational space to non-humanoid character poses in a second observation space using a model with shared latent space in a low dimensional space built based on said one to one correspondence from said key poses of said initial motion data to said non-humanoid character key poses;
wherein said model with shared latent space discovers and models space that represents a common causal structure between human and non-humanoid character motion.
2. The method according to claim 1 , further comprising optimizing statically mapped data for a non-humanoid character to create a physically correct animated motion for said non-humanoid character.
3. The method according to claim 1 , wherein said initial motion data is derived from one or more of: a motion capture system, and a key framing process.
4. The method according to claim 3 , wherein the initial motion data is human motion capture data.
5. The method according to claim 1 , wherein said selecting key poses included in said initial motion data is performed by a user operating an animation system.
6. The method according to claim 1 , wherein said non-humanoid character key poses are derived from an external device.
7. The method according to claim 1 , wherein said non-humanoid character key poses comprise key poses created by an animator collaborating with one or more actors.
8. The method according to claim 1 , wherein said statically mapping further comprises learning a static mapping function from poses of said initial motion data to said non-humanoid character poses.
9. The method according to claim 8 , wherein said model is a shared Gaussian process latent variable model.
10. The method according to claim 2 , wherein said optimizing statically mapped data for a non-humanoid character further comprises one or more of:
determining a global transformation of said non-humanoid character poses;
correcting one or more contact positions; and
optimizing physical realism of non-humanoid character motion by adjusting vertical motion to be consistent with gravity and a penalty-based contact model.
11. A computer program product comprising:
a non-transitory computer readable storage medium having computer readable program code embodied therewith, the computer readable program code comprising:
computer readable program code configured to access key poses selected from initial motion data;
computer readable program code configured to access non-humanoid character key poses which provide a one to one correspondence to selected key poses in said initial motion data; and
computer readable program code configured to statically map poses of said initial motion data in a first observational space to an non-humanoid character poses in a second observation space using a model with shared latent space in a low dimensional space built based on said one to one correspondence from said key poses of said initial motion data to said non-humanoid character key poses;
wherein said model with shared latent space discovers and models space that represents a common causal structure between human and non-humanoid character motion.
12. The computer program product according to claim 11 , further comprising computer readable program code configured to optimize statically mapped data for a non-humanoid character to create a physically correct animated motion for said non-humanoid character.
13. The computer program product according to claim 11 , wherein said initial motion data is derived from one or more of: a motion capture system, and a key framing process.
14. The computer program product according to claim 12 , wherein the initial motion data is human motion capture data.
15. The computer program product according to claim 12 , wherein said mapping further comprises learning a static mapping function from poses of said initial motion data to said non-humanoid character poses.
16. The computer program product according to claim 15 , wherein said model is a shared Gaussian process latent variable model.
17. The computer program product according to claim 12 , wherein said optimizing statically mapped data for a non-humanoid character further comprises one or more of:
determining a global transformation of said non-humanoid character poses;
correcting one or more contact positions; and
optimizing physical realism of non-humanoid character motion by adjusting motion of said non-humanoid character to be consistent with gravity and a penalty-based contact model.
18. The computer program product according to claim 17 , further comprising computer readable program code configured to match linear and angular momenta of said non-humanoid character poses to poses obtained by scaling momenta in poses of said initial motion data.
19. The computer program product according to claim 17 , wherein said optimizing physical realism of said non-humanoid character motion further comprises adjusting vertical motion of said non-humanoid character.
20. A system comprising:
at least one processor; and
a memory device operatively connected to the at least one processor;
wherein, responsive to execution of program instructions accessible to the at least one processor, the at least one processor is configured to:
access key poses selected from initial motion data;
access non-humanoid character key poses which provide a one to one correspondence to selected key poses in said initial motion data; and
statically map poses of said initial motion data in a first observational space to non-humanoid character poses in a second observation space using a model with shared latent space in a low dimensional space built based on said one to one correspondence from said key poses of said initial motion data to said non-humanoid character key poses;
wherein said model with shared latent space discovers and models space that represents a common causal structure between human and non-humanoid character motion.