Synthesizing high resolution 3D shapes from lower resolution representations for synthetic data generation systems and applications
In various examples, a deep three-dimensional (3D) conditional generative model is implemented that can synthesize high resolution 3D shapes using simple guides—such as coarse voxels, point clouds, etc.—by marrying implicit and explicit 3D representations into a hybrid 3D representation. The present approach may directly optimize for the reconstructed surface, allowing for the synthesis of finer geometric details with fewer artifacts. The systems and methods described herein may use a deformable tetrahedral grid that encodes a discretized signed distance function (SDF) and a differentiable marching tetrahedral layer that converts the implicit SDF representation to an explicit surface mesh representation. This combination allows joint optimization of the surface geometry and topology as well as generation of the hierarchy of subdivisions using reconstruction and adversarial losses defined explicitly on the surface mesh.
1. A processor comprising:
one or more circuits to:
compute, based at least on an input representation of an object, a signed distance field (SDF) at an initial grid resolution of a grid;
subdivide and deform the grid to generate an updated grid at an updated resolution;
compute an updated SDF using the SDF and the updated grid;
generate a triangular mesh using the updated SDF; and
subdivide the triangular mesh to generate a parametric surface representation of the object.
2. The processor of claim 1 , wherein the subdivision of the triangular mesh is executed using learned surface subdivision.
3. The processor of claim 1 , wherein the input representation of the object includes at least one of a voxel representation, a point cloud, or a three-dimensional (3D) scan.
4. The processor of claim 1 , wherein the updated SDF is interpolated from the SDF using one or more updated vertex positions of the updated grid.
5. The processor of claim 1 , wherein the computation of the SDF is executed by, at least in part:
computing, using a convolutional neural network, one or more first feature vectors; and
computing, using a neural network and based at least on the one or more first feature vectors, one or more SDF values for one or more vertices of the grid and one or more second feature vectors.
6. The processor of claim 1 , wherein the subdivision and the deformation of the grid is executed by, at least in part:
identifying one or more surface volumes of the grid that correspond to a surface of the object;
generate a graph corresponding to one or more vertices and one or more edges of the one or more surface volumes; and
computing, using a graph convolutional network and based at least on the graph, one or more position offsets and one or more residual SDF values for the one or more vertices.
7. The processor of claim 1 , wherein the subdivision of the grid includes selective subdivision, wherein the selective subdivision includes subdividing at least one of:
one or more first surface volumes of the grid that intersect a surface of the object; or
one or more second surface volumes that are immediately adjacent the one or more first surface volumes.
8. The process of claim 1 , wherein, the one or more circuits are to generate the parametic surface representation using a generative adversarial network (GAN).
9. The processor of claim 1 , wherein the processor is comprised in at least one of:
a control system for an autonomous or semi-autonomous machine;
a perception system for an autonomous or semi-autonomous machine;
a system for performing simulation operations;
a system for performing light transport simulation;
a system for performing collaborative content creation for 3D assets;
a system for performing deep learning operations;
a system implemented using an edge device;
a system implemented using a robot;
a system for performing conversational AI operations;
a system for generating synthetic data;
a system incorporating one or more virtual machines (VMs);
a system implemented at least partially in a data center; or
a system implemented at least partially using cloud computing resources.
10. A system comprising:
one or more processing units comprising processing circuitry to:
compute, based at least on an input representation of an object, a signed distance field (SDF) corresponding to an initial grid;
subdivide and deform the initial grid to generate an updated grid;
compute an updated SDF using the SDF and the updated grid;
generate an explicit surface representation using the updated SDF; and
subdivide the explicit surface representation to generate a parametric surface representation of the object.
11. The system of claim 10 , wherein the subdivision of the explicit surface representation is executed using learned surface subdivision.
12. The system of claim 10 , wherein the input representation of the object includes at least one of a voxel representation, a point cloud, or a three-dimensional (3D) scan.
13. The system of claim 10 , wherein the updated SDF is interpolated from the SDF using one or more updated vertex positions of the updated grid.
14. The system of claim 10 , wherein the computation of the SDF is executed by, at least in part:
computing, using a convolutional neural network, one or more first feature vectors; and
computing, using a neural network and based at least in part on the one or more first feature vectors, one or more SDF values for one or more vertices of the grid and one or more second feature vectors.
15. The system of claim 10 , wherein the subdivision and the deformation of the initial grid is executed by, at least in part:
identifying one or more surface volumes of the initial grid that correspond to a surface of the object;
generate a graph corresponding to one or more vertices and one or more edges of the one or more surface volumes; and
computing, using a graph convolutional network and based at least on the graph, one or more position offsets and one or more residual SDF values for the one or more vertices.
16. The system of claim 10 , wherein the subdivision of the grid includes selective subdivision, wherein the selective subdivision includes subdividing at least one of:
one or more first surface volumes of the grid that intersect a surface of the object; or
one or more second surface volumes that are immediately adjacent the one or more first surface volumes.
17. The system of claim 10 , wherein the one or more circuits generate the parametic surface representation using a generative adversarial network (GAN).
18. The system of claim 10 , wherein the system is comprised in at least one of:
a control system for an autonomous or semi-autonomous machine;
a perception system for an autonomous or semi-autonomous machine;
a system for performing simulation operations;
a system for performing light transport simulation;
a system for performing collaborative content creation for 3D assets;
a system for performing deep learning operations;
a system implemented using an edge device;
a system implemented using a robot;
a system for performing conversational AI operations;
a system for generating synthetic data;
a system incorporating one or more virtual machines (VMs);
a system implemented at least partially in a data center; or
a system implemented at least partially using cloud computing resources.
19. A processor comprising:
processing circuitry to:
compute, based at least on an input representation of a shape, a signed distance field (SDF) corresponding to an initial grid;
subdivide the initial grid to generate an updated grid;
compute an updated SDF using the SDF and the updated grid;
generate a mesh using the updated SDF; and
subdivide the mesh to generate a parametric surface representation of the shape.
20. The processor of claim 19 , wherein the processor is comprised in at least one of:
a control system for an autonomous or semi-autonomous machine;
a perception system for an autonomous or semi-autonomous machine;
a system for performing simulation operations;
a system for performing light transport simulation;
a system for performing collaborative content creation for 3D assets;
a system for performing deep learning operations;
a system implemented using an edge device;
a system implemented using a robot;
a system for performing conversational AI operations;
a system for generating synthetic data;
a system incorporating one or more virtual machines (VMs);
a system implemented at least partially in a data center; or
a system implemented at least partially using cloud computing resources.
21. The processor of claim 19 , wherein the computation of the SDF is executed by, at least in part:
computing, using a convolutional neural network, one or more first feature vectors; and
computing, using a neural network and based at least on the one or more first feature vectors, one or more SDF values for one or more vertices of the grid and one or more second feature vectors.