Granted Patent B2

US 12,293,463 · App. 17/661,193 · Granted May 6, 2025

Inventors: Khaled Mammou (Danville, CA); Adrian A Biagioli (Sunnyvale, CA); Deepak S Tolani (Sunnyvale, CA)

Assignee: Apple Inc.

G06T17/205G06T7/64G06T19/00G06T2207/20021G06T2210/21G06T2219/008

Abstract

Claims (87)

1. A method, performed by a computing system, of decomposing a three-dimensional representation of an object into a plurality of convex hulls, the method comprising:

instantiating a cluster priority queue in a memory of the computing system, wherein the cluster priority queue initially contains a cluster corresponding to the three-dimensional representation of the object;

computing with a processor of the computing system a concavity measure for each cluster in the cluster priority queue;

for the cluster with the highest concavity measure:

computing with the processor a cut plane that divides the cluster corresponding to the three-dimensional representation of the object into two new clusters, each of the two new clusters having a corresponding convex hull, wherein computing the cut plane includes performing a hierarchical search of potential cut planes;

removing the cluster corresponding to the three-dimensional representation of the object from the cluster priority queue; and

adding the two new clusters to the cluster priority queue;

computing a collision interaction between the object and another object based on the two new clusters; and

displaying the collision interaction.

2. The method of claim 1 wherein performing the hierarchical search of the potential cut planes includes a recursive cut plane selection procedure that minimizes concavity of all produced clusters after applying a predetermined number of successive cut planes.

3. The method of claim 1 wherein computing with the processor of the computing system the cut plane further comprises accepting user input to select or modify the cut plane.

4. The method of claim 1 wherein performing the hierarchical search of the potential cut planes further comprises:

determining a first plurality of cut plane orientations;

for each of the first plurality of cut plane orientations, computing one or more cost functions associated with the first plurality of cut plane orientations;

selecting one of the first plurality of cut plane orientations having a lowest cost as determined by the one or more cost functions;

determining a second plurality of cut plane orientations based on the selected one of the first plurality of cut plane orientations;

for each of the second plurality of cut plane orientations, computing one or more cost functions associated with the second plurality of cut plane orientations; and

selecting one of the second plurality of cut plane orientations having a lowest cost as determined by the one or more cost functions associated with the second plurality of cut plane orientations.

5. The method of claim 4 wherein the first and second pluralities of cut plane orientations are selected based on successive subdivisions of a unitary octahedron.

6. The method of claim 4 wherein the one or more cost functions associated with the first plurality of cut plane orientations, the one or more cost functions associated with the second plurality of cut plane orientations, or both include a balance cost function that prefers cut plane orientations that produce approximately equally sized clusters.

7. The method of claim 4 wherein the one or more cost functions associated with the first plurality of cut plane orientations, the one or more cost functions associated with the second plurality of cut plane orientations, or both include a cross-section cost function that prefers cut plane orientations that minimize a cross-sectional area of the intersection of the cut plane and the cluster.

8. The method of claim 4 wherein the one or more cost functions associated with the first plurality of cut plane orientations, the one or more cost functions associated with the second plurality of cut plane orientations, or both include a compactness cost function that prefers cut plane orientations that minimize a sum of surface areas of the two new clusters.

9. The method of claim 1 further comprising:

for each pair of clusters in the cluster priority queue, determining whether the pair of clusters can be merged without increasing concavity above a predetermined threshold; and

if a pair of clusters in the cluster priority queue can be merged without increasing concavity above the predetermined threshold:

merging the pair of clusters to generate a merged cluster;

removing the pair of clusters from the cluster priority queue; and

placing the merged cluster in the cluster priority queue.

10. The method of claim 9 further comprising:

accepting user input with respect to one or more cluster pairs to merge; and

for each cluster pair indicated by the user input to be merged:

merging the cluster pair to generate a user merged cluster;

removing the cluster pair from the cluster priority queue; and

placing the user merged cluster in the cluster priority queue.

11. A method, performed by a computing system, of decomposing a three-dimensional representation of an object into a plurality of convex hulls, the method comprising:

computing with a processor of the computing system a cut plane that divides a cluster stored in a memory of the computing system and corresponding to the three- dimensional representation of the object into two new clusters, each of the two new clusters having a corresponding convex hull, wherein computing the cut plane includes performing a hierarchical search of potential cut planes;

removing the cluster corresponding to the three-dimensional representation of the object from a cluster priority queue; and

adding the two new clusters to the cluster priority queue;

computing an interaction of the object based on the two new clusters; and

displaying the object based on the interaction.

12. The method of claim 11 wherein performing the hierarchical search of the potential cut planes includes a recursive cut plane selection procedure that minimizes concavity of all produced clusters after applying a predetermined number of successive cut planes.

13. The method of claim 11 wherein computing with the processor of the computing system the cut plane further comprises accepting user input to select or modify the cut plane.

14. The method of claim 11 wherein performing the hierarchical search of potential cut planes further comprises:

determining a first plurality of cut plane orientations;

for each of the first plurality of cut plane orientations, computing one or more cost functions associated with the first plurality of cut plane orientations;

selecting one of the first plurality of cut plane orientations having a lowest cost as determined by the one or more cost functions associated with the first plurality of cut plane orientations;

determining a second plurality of cut plane orientations based on the selected one of the first plurality of cut plane orientations;

for each of the second plurality of cut plane orientations, computing one or more cost functions associated with the second plurality of cut plane orientations; and

selecting one of the second plurality of cut plane orientations having a lowest cost as determined by the one or more cost functions associated with the second plurality of cut plane orientations.

15. The method of claim 14 wherein the first and second pluralities of cut plane orientations are selected based on successive subdivisions of a unitary octahedron.

16. The method of claim 14 wherein the one or more cost functions associated with the first plurality of cut plane orientations, the one or more cost functions associated with the second plurality of cut plane orientations, or both include one or more cost functions selected from the group consisting of:

a balance cost function that prefers cut plane orientations that produce approximately equally sized clusters;

a cross-section cost function that prefers cut plane orientations that minimize a cross-sectional area of the intersection of the cut plane and the cluster; and

a compactness cost function that prefers cut plane orientations that minimize a sum of surface areas of the two new clusters.

17. The method of claim 14 wherein performing the hierarchical search of the potential cut planes includes a recursive cut plane selection procedure that minimizes concavity of all produced clusters after applying a predetermined number of successive cut planes.

18. The method of claim 14 further comprising:

for each pair of clusters in a priority queue, determining whether the pair of clusters can be merged without increasing concavity above a predetermined threshold; and

if a pair of clusters in the priority queue can be merged without increasing concavity above the predetermined threshold:

merging the pair of clusters to generate a merged cluster;

removing the pair of clusters from the priority queue; and

placing the merged cluster in the priority queue.

19. The method of claim 18 further comprising:

accepting user input with respect to one or more cluster pairs to merge; and

for each cluster pair indicated by the user input to be merged:

merging the cluster pair of clusters to generate a user merged cluster;

removing the cluster pair from the priority queue; and

placing the user merged cluster in the priority queue.

20. The method of claim 11 wherein the interaction comprises a collision interaction between the object and another object.

21. A method, performed by a computing system, of performing a hierarchical search of potential cut planes for decomposing a three-dimensional representation of an object into a plurality of convex hulls, the method comprising:

determining a first plurality of cut plane orientations;

for each of the first plurality of cut plane orientations, computing one or more cost functions associated with the first plurality of cut plane orientations;

selecting one of the first plurality of cut plane orientations having a lowest cost as determined by the one or more cost functions associated with the first plurality of cut plane orientations;

determining a second plurality of cut plane orientations based on the selected one of the first plurality of cut plane orientations;

for each of the second plurality of cut plane orientations, computing one or more cost functions associated with the second plurality of cut plane orientations;

selecting one of the second plurality of cut plane orientations having a lowest cost as determined by the one or more cost functions;

determining the plurality of convex hulls based on the selected one of the second plurality of cut plane orientations;

determining an interaction of the object based on the plurality of convex hulls; and

displaying the object based on the interaction.

22. The method of claim 21 wherein performing the hierarchical search of the potential cut planes includes a recursive cut plane selection procedure that minimizes concavity of all produced clusters after applying a predetermined number of successive cut planes.

23. The method of claim 21 further comprising:

computing with a processor of the computing system a cut plane based on the selected one of the second plurality of cut plane orientations; and

dividing, based on the cut plane, a cluster stored in a memory of the computing system and corresponding to the three-dimensional representation of the object into two new clusters.

24. The method of claim 21 wherein the first and second pluralities of cut plane orientations are selected based on successive subdivisions of a unitary octahedron.

25. The method of claim 21 wherein the one or more cost functions associated with the first plurality of cut plane orientations, the one or more cost functions associated with the second plurality of cut plane orientations, or both include one or more cost functions selected from the group consisting of:

a balance cost function that prefers cut planes that produce approximately equally sized clusters;

a cross-section cost function that prefers cut plane orientations that minimize a cross-sectional area of the intersection of the cut plane orientations and an input cluster; and

a compactness cost function that prefers cut plane orientations that minimize a sum of surface areas of resulting clusters.

Assignments (1)

Continuity (1)

References Cited (106)