Method and system for motion compensation
View Patent ↗Described herein is a method and system for motion compensation. This system and method use confidence for motion vector selection. Relative cost and an estimate of noise power may be used to characterize residual values and influence confidence determination.
1. A method for motion compensation comprising:
generating a confidence metric based at least in part on a ranking of a plurality of motion vectors, wherein the ranking is based at least in part on a cost associated with each of the motion vectors, wherein the confidence metric is based at least in part on a noise probability density function; and
selecting a motion vector based at least in part on the confidence metric.
2. The method of claim 1 , wherein generating the confidence metric comprises attributing a higher confidence to a motion vector with a lower cost.
3. The method of claim 1 , further comprising adjusting the confidence metric based at least in part on a relative difference in the cost between a first motion vector and a second motion vector.
4. The method of claim 3 , wherein the relative difference in the cost is compared to a noise level for estimating a presence of video content.
5. The method of claim 3 , wherein the confidence metric is greater in the first motion vector based at least in part on the first motion vector having a lower cost than the second motion vector and the relative difference in the cost being large.
6. The method of claim 3 , wherein the confidence metric is adjusted in the first motion vector and the second motion vector based at least in part on the relative difference in cost being small.
7. The method of claim 1 , wherein the confidence metric has a vertical confidence and a horizontal confidence.
8. The method of claim 1 , wherein the cost is a weighted sum of a first measurement and a second measurement, and wherein a first expected value is subtracted from the first measurement prior to a summation.
9. The method of claim 1 , wherein generating the confidence metric further comprises eliminating motion vectors that are associated with a cost similar to a zero motion vector cost, and wherein a similarity is based at least in part on a noise level input to a video encoder.
10. The method of claim 1 , wherein the method further comprises applying the motion vector to a neighboring region of video based at least in part on the confidence metric.
11. The method of claim 1 , wherein following an application of a motion vector, the method further comprises confirming confidence based at least in part on a comparison of a residual value to a noise level.
12. A system for motion compensation, said system comprising:
a first circuit configured to generate a confidence metric based on a ranking of a plurality of motion vectors, wherein the ranking is based on a cost associated with each of the motion vectors, wherein generating the confidence metric comprises determining a noise probability density function; and
a second circuit configured to select, based at least in part on the confidence metric, a motion vector from the plurality of motion vectors.
13. The system of claim 12 , wherein the first circuit attributes a higher confidence to a motion vector with a lower cost.
14. The system of claim 12 , wherein the first circuit adjusts the confidence metric based at least in part on a relative difference in cost between a first motion vector and a second motion vector.
15. The system of claim 14 , wherein the first circuit compares the relative difference in cost to a noise level for estimating a presence of video content.
16. The system of claim 14 , wherein the first circuit increases confidence in the first motion vector based at least in part on the first motion vector having a lower cost than the second motion vector and the relative difference in cost being large.
17. The system of claim 14 , wherein the first circuit decreases confidence in the first motion vector and the second motion vector based at least in part on the relative difference in cost being small.
18. The system of claim 12 , wherein the confidence metric has a vertical confidence and a horizontal confidence.
19. The system of claim 12 , wherein the first circuit generates the cost with a weighted sum of a first measurement and a second measurement, and wherein a first expected value is subtracted from the first measurement prior to a summation.
20. The system of claim 12 , wherein the first circuit is further configured to eliminate motion vectors that are associated with a cost similar to a zero motion vector cost, wherein a similarity is based at least in part on a noise level input to a video encoder.
21. The system of claim 12 , wherein the second circuit is further configured to apply the motion vector to a neighboring region of video based at least in part on the confidence metric.
22. The system of claim 12 , wherein the first circuit is configured to receive an output of the second circuit to confirm confidence based at least in part on a comparison of a residual value to a noise level.
23. A method for motion compensation comprising:
determining a ranking of a plurality of motion vectors based at least in part on a cost metric associated with each of the motion vectors;
generating a confidence metric based at least in part on the ranking of the motion vectors, the confidence metric generated utilizing a noise probability density function;
adjusting the confidence metric associated with a first one of the motion vectors in response to a similarity of the cost metric associated with a first one of the motion vectors is to the cost metric associated with a second one of the motion vectors, the similarity based at least in part on a measured noise; and
selecting at least one of the motion vectors based at least in part on the confidence metric associated with the at least one of the motion vectors.