Inferred skeletal structure for practical 3D assets
Some implementations relate to methods, systems, and computer-readable media for inferring a skeletal structure to create a practical 3D asset. A user may create an avatar using different parts from different parent models. Upon creation, a data model for the avatar is examined to determine joint placement and infer an overall skeletal structure. Thereafter, different skinning techniques may be used to deform the avatar skin during animations based on the joint placement and movement.
1 . A computer-implemented method, comprising:
receiving an avatar data model for a three-dimensional (3D) avatar, the avatar data model comprising at least a first body part from a first parent model, and a second body part from a second parent model, wherein the first parent model is a first 3D model, and wherein the second parent model is a second 3D model;
creating a graph representation for the avatar data model, wherein the graph representation begins at a selected root point for the avatar data model and extends to terminal portions of the avatar data model, wherein the graph representation comprises a plurality of nodes labeled according to corresponding individual body parts of the avatar data model;
traversing the graph representation to determine a plurality of joint locations, wherein the joint locations each represent a respective location where a joint is to be placed, wherein the joint locations are determined based on intersections of different body parts, descriptions of different body parts, or connections between different body parts, and wherein at least one joint location of the joint locations represents an intersection between the first body part and the second body part;
inferring a skeletal structure between the plurality of joint locations such that,
for all joints, a skeletal portion extends between a first joint and a second joint, the skeletal portion rotates by motion of one or both of the first joint and the second joint, and
for the at least one joint location, at least one skeletal portion extends from the at least one joint location based on a scale of a corresponding skeletal portion of the second 3D model; and
combining joints at each of the plurality of joint locations, the inferred skeletal structure, and the avatar data model to create a practical 3D asset.
2 . The computer-implemented method of claim 1 , further comprising:
animating the practical 3D asset in a virtual environment of a virtual experience.
3 . The computer-implemented method of claim 2 , wherein animating the practical 3D asset comprises utilizing the inferred skeletal structure for skin deformations during animation of the practical 3D asset.
4 . The computer-implemented method of claim 1 , further comprising creating a weight map for individually placed joints, wherein the weight map controls a degree of movement of each joint based upon the inferred skeletal structure.
5 . The computer-implemented method of claim 1 , wherein the graph representation further comprises a plurality of edges, wherein each edge of the plurality of edges extends between at least two nodes of the plurality of nodes, and wherein each edge of the plurality of edges represents a respective joint location, wherein each node that is linked to two or more edges has one or more virtual bones inferred between the two or more edges to manage positions of non-distal nodes of the plurality of nodes.
6 . The computer-implemented method of claim 1 , wherein each joint combined at the plurality of joint locations comprises either an animated joint or a fixed joint.
7 . The computer-implemented method of claim 1 , further comprising skinning the practical 3D asset based upon at least the inferred skeletal structure.
8 . The computer-implemented method of claim 1 , wherein traversing the graph representation comprises one or more of: a depth first search or a breadth first search.
9 . The computer-implemented method of claim 1 , wherein the graph representation comprises a weighted graph having weights assigned to edges based upon a scale of individual body parts of the avatar data model, wherein the scale informs inference of the inferred skeletal structure.
10 . The computer-implemented method of claim 1 , wherein one or more of the first parent model or the second parent model are user-generated content of a virtual experience platform.
11 . A system, comprising:
a memory with instructions stored thereon; and
a processing device, coupled to the memory and operable to access the memory, wherein the instructions when executed by the processing device, cause the processing device to perform operations including:
receiving an avatar data model for a three-dimensional (3D) avatar, the avatar data model comprising at least a first body part from a first parent model, and a second body part from a second parent model, wherein the first parent model is a first 3D model, and wherein the second parent model is a second 3D model;
creating a graph representation for the avatar data model, wherein the graph representation begins at a selected root point for the avatar data model and extends to terminal portions of the avatar data model, wherein the graph representation comprises a plurality of nodes labeled according to corresponding individual body parts of the avatar data model;
traversing the graph representation to determine a plurality of joint locations, wherein the joint locations each represent a respective location where a joint is to be placed, wherein the joint locations are determined based on intersections of different body parts, descriptions of different body parts, or connections between different body parts, and wherein at least one joint location of the joint locations represents an intersection between the first body part and the second body part;
inferring a skeletal structure between the plurality of joint locations such that,
for all joints, a skeletal portion extends between a first joint and a second joint, the skeletal portion rotates by motion of one or both of the first joint and the second joint, and
for the at least one joint location, at least one skeletal portion extends from the at least one joint location based on a scale of a corresponding skeletal portion of the second 3D model; and
combining joints at each of the plurality of joint locations, the inferred skeletal structure, and the avatar data model to create a practical 3D asset.
12 . The system of claim 11 , wherein the operations further comprise animating the practical 3D asset in a virtual environment of a virtual experience.
13 . The system of claim 12 , wherein animating the practical 3D asset comprises utilizing the inferred skeletal structure for skin deformations during animation of the practical 3D asset.
14 . The system of claim 11 , wherein the graph representation further comprises a plurality of edges, wherein each edge of the plurality of edges extends between at least two nodes of the plurality of nodes, and wherein each edge of the plurality of edges represents a respective joint location, wherein each node that is linked to two or more edges has one or more virtual bones inferred between the two or more edges to manage positions of non-distal nodes of the plurality of nodes.
15 . The system of claim 14 , wherein the virtual bones are scaled based upon a customized version of the avatar data model and the corresponding individual body parts.
16 . The system of claim 11 , wherein each joint combined at the plurality of joint locations comprises either an animated joint or a fixed joint.
17 . The system of claim 11 , wherein the operations further comprise skinning the practical 3D asset based upon at least the inferred skeletal structure.
18 . The system of claim 11 , wherein traversing the graph representation comprises one or more of: a depth first search or a breadth first search.
19 . The system of claim 11 , wherein the graph representation comprises a weighted graph having weights assigned to edges based upon a scale of individual body parts of the avatar data model.
20 . A non-transitory computer-readable medium with instructions stored thereon that, responsive to execution by a processing device, causes the processing device to perform operations comprising:
receiving an avatar data model for a three-dimensional (3D) avatar, the avatar data model comprising at least a first body part from a first parent model, and a second body part from a second parent model, wherein the first parent model is a first 3D model, and wherein the second parent model is a second 3D model;
creating a graph representation for the avatar data model, wherein the graph representation begins at a selected root point for the avatar data model and extends to terminal portions of the avatar data model, wherein the graph representation comprises a plurality of nodes labeled according to corresponding individual body parts of the avatar data model;
traversing the graph representation to determine a plurality of joint locations, wherein the joint locations each represent a respective location where a joint is to be placed, wherein the joint locations are determined based on intersections of different body parts, descriptions of different body parts, or connections between different body parts, and wherein at least one joint location of the joint locations represents an intersection between the first body part and the second body part;
inferring a skeletal structure between the plurality of joint locations such that,
for all joints, a skeletal portion extends between a first joint and a second joint, the skeletal portion rotates by motion of one or both of the first joint and the second joint, and
for the at least one joint location, at least one skeletal portion extends from the at least one joint location based on a scale of a corresponding skeletal portion of the second 3D model; and
combining joints at each of the plurality of joint locations, the inferred skeletal structure, and the avatar data model to create a practical 3D asset.