Systems and methods for optimizing order of image transformations

Patent metadata

Field	Value
Publication number	US-9361659-B2
Application number	US-201213677206-A
Country	US
Kind code	B2
Filing date	Nov 14, 2012
Priority date	Nov 14, 2012
Publication date	Jun 7, 2016
Grant date	Jun 7, 2016

Abstract

Techniques to optimize the ordering of image transformations. Ordering cost functions are applied to possible transformation orderings of image transformations of a user-selected transformation ordering. Computing costs are calculated based on the cost functions. An optimal transformation ordering is selected from the possible transformation orderings based on the computing costs. In one embodiment, a first transformation and a second transformation of the user-selected transformation ordering are swapped. In one embodiment, at least one of the ordering cost functions is based on a per-pixel cost value. In one embodiment, a fidelity loss threshold representing an acceptable level of difference between a first image resulting from the optimal transformation ordering and a second image resulting from the user-selected transformation ordering is determined. In one embodiment, selection of the optimal ordering transformation is based on satisfaction of the fidelity loss threshold.

First claim

The invention claimed is:

1. A computer implemented method comprising:
applying, by a computer system, one or more ordering cost functions to each possible transformation ordering of a plurality of image transformations of a user-selected transformation ordering, each ordering cost function comprising a plurality of transformation cost functions, and each transformation cost function being associated with an image transformation and based on a per-pixel cost value;
determining, by the computer system, computing costs for each of the one or more applied ordering cost functions; and
selecting, by the computer system, an optimal transformation ordering of the plurality of image transformations based on the determined computing costs.

2. The computer implemented method of claim 1 , further comprising swapping a first transformation and a second transformation of the user-selected transformation ordering.

3. The computer implemented method of claim 2 , further comprising adjusting at least one of a first parameter associated with the first transformation and a second parameter associated with the second transformation.

4. The computer implemented method of claim 3 , wherein the adjusting the at least one of the first parameter associated with the first transformation and the second parameter associated with the second transformation is based on at least one heuristic.

5. The computer implemented method of claim 2 , wherein the swapping the first transformation and the second transformation of the user-selected transformation ordering comprises substituting an alternate technique for at least one of the first transformation and the second transformation.

6. The computer implemented method of claim 1 , wherein the at least one of the ordering cost functions is further based on a function form.

7. The computer implemented method of claim 6 , wherein at least one of the per-pixel cost value and the function form is based on hardware capabilities.

8. The computer implemented method of claim 1 , further comprising identifying collapsing transformations from the image transformations.

9. The computer implemented method of claim 1 , further comprising identifying combining transformations from the image transformations.

10. The computer implemented method of claim 1 , further comprising identifying a cost savings feature based on parameters associated with at least one of the image transformations, wherein at least one of the ordering cost functions is based on the cost savings feature.

11. The computer implemented method of claim 10 , wherein the cost savings feature is based on use of a technique for executing at least one of the image transformations.

12. The computer implemented method of claim 10 , wherein the cost savings feature is based on use of a hardware component.

13. The computer implemented method of claim 1 , further comprising determining a fidelity loss threshold representing an acceptable level of difference between a first image resulting from the optimal transformation ordering and a second image resulting from the user-selected transformation ordering.

14. The computer implemented method of claim 13 , wherein the selecting the optimal ordering transformation is based on satisfaction of the fidelity loss threshold.

15. The computer implemented method of claim 1 , wherein the image transformations comprise at least one of a crop, a rotate, and a shrink.

16. The computer implemented method of claim 1 , wherein the determining computing costs comprises:
determining at least one sequence cost for at least one sequence of the possible transformation orderings; and
storing the at least one sequence cost.

17. The computer implemented method of claim 1 , further comprising identifying at least one constraint associated with at least one of the image transformations of the user-selected transformation ordering, wherein the selecting an optimal transformation ordering is based on the at least one constraint.

18. A system comprising:
at least one processor; and
a memory storing instructions configured to instruct the at least one processor to perform:
applying, by a computer system, one or more ordering cost functions to each possible transformation ordering of a plurality of image transformations of a user-selected transformation ordering, each ordering cost function comprising a plurality of transformation cost functions, and each transformation cost function being associated with an image transformation and based on a per-pixel cost value;
determining computing costs for each of the one or more applied ordering cost functions; and
selecting an optimal transformation ordering of the plurality of user-selected image transformations based on the determined computing costs.

19. The system of claim 18 , wherein the instructions are further configured to instruct the at least one processor to perform swapping a first transformation and a second transformation of the user-selected transformation ordering.

20. The system of claim 19 , wherein the instructions are further configured to instruct the at least one processor to perform adjusting at least one of a first parameter associated with the first transformation and a second parameter associated with the second transformation.

21. The system of claim 20 , wherein the adjusting the at least one of the first parameter associated with the first transformation and the second parameter associated with the second transformation is based on at least one heuristic.

22. The system of claim 19 , wherein the swapping the first transformation and the second transformation of the user-selected transformation ordering comprises substituting an alternate technique for at least one of the first transformation and the second transformation.

23. The system of claim 18 , wherein the at least one of the ordering cost functions is further based on a function form.

24. The system of claim 23 , wherein at least one of the per-pixel cost value and the function form is based on hardware capabilities.

25. The system of claim 18 , wherein the instructions are further configured to instruct the at least one processor to perform identifying collapsing transformations from the image transformations.

26. The system of claim 18 , wherein the instructions are further configured to instruct the at least one processor to perform identifying combining transformations from the image transformations.

27. The system of claim 18 , wherein the instructions are further configured to instruct the at least one processor to perform identifying a cost savings feature based on parameters associated with at least one of the image transformations, wherein at least one of the ordering cost functions is based on the cost savings feature.

28. The system of claim 27 , wherein the cost savings feature is based on use of a technique for executing at least one of the image transformations.

29. The system of claim 27 , wherein the cost savings feature is based on use of a hardware component.

30. The system of claim 18 , wherein the instructions are further configured to instruct the at least one processor to perform determining a fidelity loss threshold representing an acceptable level of difference between a first image resulting from the optimal transformation ordering and a second image resulting from the user-selected trans

Assignees

• Lerios Apostolos
• Facebook Inc

Inventors

• Lerios Apostolos

Classifications

• G06T1/20

  Processor architectures; Processor configuration, e.g. pipelining · CPC title

• G06T11/60

  Creating or editing images; Combining images with text · CPC title

• G06T2210/22

  Cropping · CPC title

• G06T1/00 Primary

  General purpose image data processing · CPC title

• G06F9/3851

  from multiple instruction streams, e.g. multistreaming · CPC title