Methods and systems for interpolation of disparate inputs
Systems and methods are provided for interpolation of disparate inputs. A radial basis function neural network (RBFNN) may be used to interpolate the pose of a digital character. Input parameters to the RBFNN may be separated by data type (e.g. angular vs. linear) and manipulated within the RBFNN by distance functions specific to the data type (e.g. use an angular distance function for the angular input data). A weight may be applied to each distance to compensate for input data representing different variables (e.g. clavicle vs. shoulder). The output parameters of the RBFNN may be a set of independent values, which may be combined into combination values (e.g. representing x, y, z, w angular value in SO(3) space).
1. A method, comprising:
receiving at least one input parameter to a radial basis function neural network (RBFNN);
calculating a set of distances from one or more distance functions within the RBFNN; and
applying a set of distance weights to the set of distances calculated from the one or more distance functions within the RBFNN;
wherein the set of distance weights comprises one or more distance weights;
wherein the set of weights comprises a distance weight for each input parameter to the RBFNN; and
wherein the at least one input parameter represents a low-order digital character rig.
2. The method of claim 1 , comprising defining at least one distance weight from the set of distance weights utilizing user input.
3. The method of claim 1 , comprising defining at least one distance weight from the set of distance weights utilizing the RBFNN.
4. The method of claim 1 , further comprising separating the input parameters into:
a first input data type comprising data representing a linear parameter; and
a second input data type comprising data representing an angular parameter.
5. The method of claim 1 , wherein the one or more distance functions comprise at least one angular distance function.
6. The method of claim 5 , wherein the angular distance function is in special orthogonal group in three dimensions (SO(3)) space.
7. The method of claim 5 , wherein the angular distance function is in quaternion space.
8. The method of claim 1 , wherein the one or more distance functions comprise at least one linear distance function.
9. The method of claim 8 , wherein the linear distance function is a Euclidean distance function.
10. The method of claim 1 , comprising interpolating a pose for a digital character utilizing the RBFNN.
11. The method of claim 1 , comprising calculating a plurality of output parameters that are independent values utilizing the RBFNN.
12. The method of claim 11 , further comprising grouping at least two of the plurality of independent values into a combination value.
13. The method of claim 12 , wherein the combination value represents a quaternion vector comprising x, y, z, and w.
14. The method of claim 1 , wherein calculating comprises calculating, utilizing the RBFNN, output parameters representing a high-order digital character rig.
15. A method, comprising:
receiving at least one input parameter to a radial basis function neural network (RBFNN);
calculating a set of distances from one or more distance functions within the RBFNN; and
applying a set of distance weights to the set of distances calculated from the one or more distance functions within the RBFNN;
wherein the set of distance weights comprises one or more distance weights; and
wherein the set of weights comprises a distance weight for each input parameter to the RBFNN,
wherein calculating comprises calculating, utilizing the RBFNN, output parameters representing a high-order digital character rig.
16. The method of claim 12 , wherein the at least one input parameter represents a low-order digital character rig.
17. A system, comprising:
a radial basis function neural network (RBFNN) implemented on one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to:
receive at least one input parameter from a set of input parameters into the RBFNN;
calculate a set of distances from one or more distance functions within the RBFNN; and
apply a set of distance weights to the set of distances calculated from the one or more distances within the RBFNN, wherein the set of distance weights comprise a distance weight for each input parameter to the RBFNN;
wherein the one or more input parameters represent a low-order digital character rig; and
wherein the RBFNN is configured to calculate output parameters that represent a high-order digital character rig.
18. The system of claim 17 , wherein the at least one distance weight from the set of distance weights is defined by the RBFNN.
19. The system of claim 17 , further comprising separating the input parameters into:
a first input data type comprising data representing a linear parameter; and
a second input data type comprising data representing an angular parameter.
20. The method of claim 17 , wherein the RBFNN is configured to interpolate a pose for a digital character.