Transforming a three-dimensional virtual model to a manipulatable format

1. A three-dimensional (3D) content refactoring system, comprising:

electronic memory to store a 3D donor asset that does not adhere to a target content standard; and

one or more processing units to:

generate a map that correlates landmarks on the 3D donor asset to corresponding landmarks on a 3D target asset, wherein the 3D target asset adheres to the target content standard;

generate a blendshape based on positional differences between vertices of the 3D target asset and a shape of the 3D donor asset using the map, wherein to generate the blendshape the one or more processing units:

generate a clone of the 3D target asset to generate a dummy mesh;

translate vertices of the dummy mesh to match the donor mesh at the landmark points and vertices surrounding the landmark points;

cut the dummy mesh in half vertically;

refine a shape of the half of the dummy mesh to more precisely match the 3D donor asset; and

mirror the half of the dummy mesh to generate the blendshape; and

drive a deformation in the 3D target asset using the blendshape to generate a new 3D asset that looks like the 3D donor asset and adheres to the target content standard.

2. The system of claim 1 , wherein the 3D donor asset and the 3D target asset are humanoid models.

3. The system of claim 1 , wherein the one or more processing units are further to identify the corresponding landmarks based on similar features or anatomical landmarks of the 3D donor asset and the 3D target asset.

4. The system of claim 3 , wherein the corresponding landmarks include vertices associated with a nose, a corner of a mouth, fingertips, knuckles, a chin, ears, and eyes.

5. The system of claim 1 , wherein the target content standard comprises specific rules for 3D models, blendshapes, UV and texture mapping, rig modification, and following assets.

6. The system of claim 1 , wherein the one or more processing units are further to identify areas of the 3D donor asset that are asymmetrical and modify the new 3D asset to include asymmetries.

7. The system of claim 1 , wherein the one or more processing units are further to:

determine a texture from the 3D donor asset; and

project the texture onto the new 3D asset.

8. The system of claim 1 , wherein the one or more processing units are further to:

generate new normal maps for the new 3D asset,

wherein when the 3D donor asset comprises donor normal maps, and the one or more processing units are to:

translate the donor normal maps from tangent space to world space;

recalculate a tangent space normal map to a UV map of the 3D target asset; and

wherein if the 3D donor asset does not comprise the donor normal map, the one or more processing units are to bake the new 3D asset using the 3D donor asset to generate the new normal maps for the new 3D asset.

9. A method for refactoring three-dimensional (3D) assets, the method comprising:

receiving a 3D donor asset that does not adhere to a target content standard;

mapping landmarks on the 3D donor asset to corresponding landmarks on a 3D target asset, wherein the 3D target asset adheres to the target content standard;

generating a blendshape based on positional differences between vertices of the 3D target asset and a shape of the 3D donor asset using a map generated from mapping the landmarks on the 3D donor asset to the corresponding landmarks on a 3D target asset, wherein generating the blendshape comprises:

generating a clone of the 3D target asset to generate a dummy mesh; and

translating vertices of the dummy mesh to match the donor mesh at the landmark points and vertices surrounding the landmark points;

cutting the dummy mesh in half vertically;

refining a shape of the half of the dummy mesh to more precisely match the 3D donor asset; and

mirroring the half of the dummy mesh to generate the blendshape; and

driving a deformation in the 3D target asset using the blendshape to generate a new 3D asset that looks like the 3D donor asset and adheres to the target content standard.

10. The method of claim 9 , wherein the 3D donor asset and the 3D target asset are humanoid models.

11. The method of claim 9 , further comprising identifying the corresponding landmarks based on similar features or anatomical landmarks of the 3D donor asset and the 3D target asset.

12. The method of claim 11 , wherein the corresponding landmarks include vertices associated with a nose, a corner of a mouth, fingertips, knuckles, a chin, ears, and eyes.

13. The method of claim 9 , wherein the target content standard comprises specific rules for 3D models, blendshapes, UV and texture mapping, rig modification, and following assets.

14. The method of claim 9 , further comprising:

identifying areas of the 3D donor asset that are asymmetrical; and

modifying the new 3D asset to include the asymmetries.

15. The method of claim 9 , further comprising:

determining a texture from the 3D donor asset; and

projecting the texture onto the new 3D asset.

16. The method of claim 9 , further comprising:

generating new normal maps for the new 3D asset,

wherein when the 3D donor asset comprises donor normal maps, generating the new normal maps comprises:

translating the donor normal maps from tangent space to world space,

recalculating a tangent space normal map to a UV map of the 3D target asset, and

wherein if the 3D donor asset does not comprise the donor normal map, generating the new normal maps comprising baking the new 3D asset using the 3D donor asset to generate the new normal maps for the new 3D asset.