Systems and methods for client-controlled adaptive streaming of three-dimensional content

1. A method comprising:

generating a three-dimensional (3D) model at a plurality of fidelities, wherein the 3D model at a first fidelity of the plurality of fidelities represents a scene or object with a first number of 3D primitives and the 3D model at a second fidelity of the plurality of fidelities represents the scene or object with a second number of 3D primitives;

partitioning the 3D model into a plurality of volumetric spatial units;

generating a manifest comprising an index for each volumetric spatial unit of the plurality of volumetric spatial units and an identifier for each fidelity of the plurality of fidelities; and

presenting 3D primitives in a first set of the plurality of volumetric spatial units at a first fidelity and 3D primitives in a second set of the plurality of volumetric spatial units at a second fidelity in response to a request comprising indices of the first set of volumetric spatial units with a first identifier for the first fidelity and indices of the second set of volumetric spatial units with a second identifier for the second fidelity.

2. The method of claim 1 further comprising:

receiving the request; and

providing the manifest in response to the request.

3. The method of claim 1 , wherein presenting the 3D primitives comprises:

streaming the first number of 3D primitives that are defined in each of the first set of volumetric spatial units; and

streaming the second number of 3D primitives that are defined in each of the second set of volumetric spatial units.

4. The method of claim 1 , wherein presenting the 3D primitives comprises:

representing parts of the 3D model within the first set of volumetric spatial units with a first level-of-detail; and

representing parts of the 3D model within the second set of volumetric spatial units with a second level-of-detail that is greater than the first level-of-detail, and wherein the parts of the 3D model within the second set of volumetric spatial units are represented with more 3D primitives than the parts of the 3D model within the first set of volumetric spatial units.

5. The method of claim 1 , wherein partitioning the 3D model comprises:

defining each volumetric spatial unit of the plurality of volumetric spatial units as a non-overlapping region that spans a different volume or part of the 3D model.

6. The method of claim 1 , wherein generating the 3D model comprises:

generating the 3D model at the first fidelity with the first number of 3D primitives; and

generating the 3D model at the second fidelity with the second number of 3D primitives added to the first number of 3D primitives.

7. The method of claim 1 , wherein generating the 3D model comprises:

associating a first identifier for the first fidelity to the first number of 3D primitives; and

associating a second identifier for the second fidelity to the second number of 3D primitives.

8. The method of claim 1 , presenting the 3D primitives comprises:

presenting the 3D primitives in the first set of the plurality of volumetric spatial units that are associated with the first identifier in response to the request; and

presenting the 3D primitives in the second set of the plurality of volumetric spatial units that are associated with the first identifier or the second identifier in response to the request.

9. The method of claim 1 , wherein generating the 3D model comprises:

generating a first set of Gaussian splats that represent the scene or object at the first fidelity; and

generating a second set of Gaussian splats with a greater density and that are smaller in size than the first set of Gaussian splats to represent the scene or object at the second fidelity.

10. The method of claim 1 , wherein generating the 3D model comprises:

generating a first set of Gaussian splats that represent the scene or object at the first fidelity; and

generating a second set of Gaussian splats that are rendered in combination with the first set of Gaussian splats to represent the scene or object at the second fidelity.

11. The method of claim 1 wherein generating the 3D model comprises:

generating a first set of points of a point cloud that represent the scene or object at a first level-of-detail corresponding to the first fidelity; and

generating a second set of points of the point cloud that are rendered in combination with the first set of points to represent the scene or object at a second level-of-detail corresponding to the second fidelity.

12. The method of claim 1 further comprising:

generating a hierarchical representation comprising a first layer of nodes corresponding to the plurality of volumetric spatial units, a second layer of nodes linked to the first layer of nodes that are associated with the first number of 3D primitives representing the 3D model at the first fidelity, and a third layer of nodes linked to the second layer of nodes that are associated with the second number of 3D primitives.

13. The method of claim 12 , wherein generating the hierarchical representation comprises:

linking nodes from the second layer of nodes that are associated with 3D primitives at a first fidelity in a specific volumetric spatial unit to a particular node of the first layer of nodes that represents the specific volumetric spatial unit.

14. A three-dimensional (3D) streaming system comprising:

one or more hardware processors configured to:

generate a 3D model at a plurality of fidelities, wherein the 3D model at a first fidelity of the plurality of fidelities represents a scene or object with a first number of 3D primitives and the 3D model at a second fidelity of the plurality of fidelities represents the scene or object with a second number of 3D primitives;

partition the 3D model into a plurality of volumetric spatial units;

generate a manifest comprising an index for each volumetric spatial unit of the plurality of volumetric spatial units and an identifier for each fidelity of the plurality of fidelities; and

present 3D primitives in a first set of the plurality of volumetric spatial units at a first fidelity and 3D primitives in a second set of the plurality of volumetric spatial units at a second fidelity in response to a request comprising indices of the first set of volumetric spatial units with a first identifier for the first fidelity and indices of the second set of volumetric spatial units with a second identifier for the second fidelity.

15. The 3D streaming system of claim 14 , wherein the one or more hardware processors are further configured to:

receive the request; and

provide the manifest in response to the request.

16. The 3D streaming system of claim 14 , wherein presenting the 3D primitives comprises:

streaming the first number of 3D primitives that are defined in each of the first set of volumetric spatial units; and

streaming the second number of 3D primitives that are defined in each of the second set of volumetric spatial units.

17. The 3D streaming system of claim 14 , wherein presenting the 3D primitives comprises:

representing parts of the 3D model within the first set of volumetric spatial units with a first level-of-detail; and

representing parts of the 3D model within the second set of volumetric spatial units with a second level-of-detail that is greater than the first level-of-detail, and wherein the parts of the 3D model within the second set of volumetric spatial units are represented with more 3D primitives than the parts of the 3D model within the first set of volumetric spatial units.

18. The 3D streaming system of claim 14 , wherein partitioning the 3D model comprises:

defining each volumetric spatial unit of the plurality of volumetric spatial units as a non-overlapping region that spans a different volume or part of the 3D model.

19. The 3D streaming system of claim 14 , wherein generating the 3D model comprises:

generating the 3D model at the first fidelity with the first number of 3D primitives; and

generating the 3D model at the second fidelity with the second number of 3D primitives added to the first number of 3D primitives.

20. A non-transitory computer-readable medium storing instructions that, when executed by one or more processors of a three-dimensional (3D) streaming system, cause the 3D streaming system to perform operations comprising:

generating a three-dimensional (3D) model at a plurality of fidelities, wherein the 3D model at a first fidelity of the plurality of fidelities represents a scene or object with a first number of 3D primitives and the 3D model at a second fidelity of the plurality of fidelities represents the scene or object with a second number of 3D primitives;

partitioning the 3D model into a plurality of volumetric spatial units;

generating a manifest comprising an index for each volumetric spatial unit of the plurality of volumetric spatial units and an identifier for each fidelity of the plurality of fidelities; and

presenting 3D primitives in a first set of the plurality of volumetric spatial units at a first fidelity and 3D primitives in a second set of the plurality of volumetric spatial units at a second fidelity in response to a request comprising indices of the first set of volumetric spatial units with a first identifier for the first fidelity and indices of the second set of volumetric spatial units with a second identifier for the second fidelity.