Image regularization and retargeting system
Systems and methods for image retargeting are provided. Image data may be acquired that includes motion capture data indicative of motion of a plurality of markers disposed on a surface of a first subject. Each of the markers may be associated with a respective location on the first subject. A plurality of blendshapes may be calculated for the motion capture data based on a configuration of the markers. An error function may be identified for the plurality of blendshapes, and it may be determined that the plurality of blendshapes can be used to retarget a second subject based on the error function. The plurality of blendshapes may then be applied to a second subject to generate a new animation.
1. A computer-implemented method for depiction of complex geometries, the method comprising:
acquiring a motion capture data sequence of a subject, the motion capture data including body position tracking data of the subject having at least one body surface;
modeling the motion capture data as an animation model that includes a set of a plurality of corresponding blendshapes associated with regularization parameters, wherein the blendshapes include at least one non-linear term and at least one linear term;
adjusting the regularization parameters based on one or more physiological factors associated with activation of neural pathways and non-physiological factors, wherein a number of the blendshapes in the set is reduced by eliminating at least one blendshape associated with an erroneous blendshape combination, wherein the regularization parameters are adjusted based on an energy function representing the activation of neural pathways that includes a threshold action potential; and
generating a representation of the at least one body surface of the subject based on the reduced set of blendshapes within the animation model in accordance with the motion capture data and adjusted regularization parameters.
2. The method of claim 1 , wherein the non-linear term represents an incremental shape.
3. The method of claim 1 , wherein the non-linear term is a sum of a non- linear effect of combining one or more of the plurality of blendshapes.
4. The method of claim 3 , wherein the non-linear term reduces a number of blendshapes used for modeling.
5. The method of claim 1 , wherein the physiological factors include human or non-human bodies or parts.
6. The method of claim 1 , wherein the physiological factors includes characteristics of a muscle or collection of muscles.
7. The method of claim 1 , wherein the regularization parameters are adjusted to reduce activation of the neural pathways.
8. The method of claim 1 , wherein the non-physiological factors include a user preference.
9. The method of claim 1 , wherein generating the representation includes evaluating an error function comprising a weighted combination of the regularization parameters.
10. The method of claim 9 , further comprising updating a temporal dependence of weights used for the weighted combination.
11. The method of claim 9 , wherein the weighted combination of the regularization parameters are based on relative probability of activation.
12. The method of claim 9 , wherein the regularization parameters are adjusted by constraining the weighted combination of the regularization parameters to be less than or equal to one.
13. The method of claim 9 , wherein the blendshapes include at least one non-linear term expressed as an algebraic expansion of the weighted combination of the regularization parameters.
14. The method of claim 13 , wherein the algebraic expansion is a Taylor expansion.
15. The method of claim 1 , wherein the regularization parameters are adjusted by introducing a corrective delta shape by a user.
16. The method of claim 1 , wherein the regularization parameters are adjusted by introducing a temporal smoothing function based on a rate of reaction of the blendshapes.
17. The method of claim 1 , wherein the regularization parameters are adjusted to have a temporal dependence expressed as a response curve.
18. The method of claim 1 , wherein generating the representation includes utilizing an energy equation that reduces energy required to activate muscles.
19. The method of claim 1 , further comprising storing a lookup table for covariance between muscles, wherein the regularization parameters are adjusted to prevent co-activation of opposing shapes of the muscles.
20. A system for depiction of complex geometries, the system comprising:
memory that stores instructions; and
at least one processor that executes the instructions stored in memory, wherein execution of the instructions by the at least one processor:
acquires a motion capture data sequence of a subject, the motion capture data including body position tracking data of the subject having at least one body surface;
models the motion capture data as an animation model that includes a set of a plurality of corresponding blendshapes associated with regularization parameters, wherein the blendshapes include at least one non-linear term and at least one linear term;
adjusts the regularization parameters based on one or more physiological factors associated with activation of neural pathways and non-physiological factors, wherein a number of the blendshapes in the set is reduced by eliminating at least one blendshape associated with an erroneous blendshape combination, wherein the regularization parameters are adjusted based on an energy function representing the activation of neural pathways that includes a threshold action potential; and
generates a representation of the at least one body surface of the subject based on the reduced set of blendshapes within the animation model in accordance with the motion capture data and adjusted regularization parameters.
21. A non-transitory computer-readable storage medium, having embodied thereon a program executable by a processor to perform a method for depiction of complex geometries, the method comprising:
acquiring a motion capture data sequence of a subject, the motion capture data including body position tracking data of the subject having at least one body surface;
modeling the motion capture data as an animation model that includes a set of a plurality of corresponding blendshapes associated with regularization parameters, wherein the blendshapes include at least one non-linear term and at least one linear term;
adjusting the regularization parameters based on one or more physiological factors associated with activation of neural pathways and non-physiological factors, wherein a number of the blendshapes in the set is reduced by eliminating at least one blendshape associated with an erroneous blendshape combination, wherein the regularization parameters are adjusted based on an energy function representing the activation of neural pathways that includes a threshold action potential; and
generating a representation of the at least one body surface of the subject based on the reduced set of blendshapes within the animation model in accordance with the motion capture data and adjusted regularization parameters.