Method and system for video evaluation in the presence of cross-chroma interference
View Patent ↗Described herein is a method and system for video sequence evaluation in the presence of cross-chroma interference. When the luma and chroma components in an analog video sequence are separated, the luma component may add interference to the chroma component. This is called cross-chroma interference. In the evaluation of the video sequence, the cross-chroma interference may be interpreted as motion, noise, or other artifacts. By reducing the cross-chroma interference during the evaluation, motion and noise can be identified with greater probability. The proper identification can aid motion estimators used in video coding and motion compensated temporal filtering, and it can aid video filters that do not utilize motion compensation.
1. A method for video evaluation in the presence of cross-chroma interference, said method comprising:
generating a predicted picture based on a reference picture with a circuit;
generating a residual value based on a plurality of chroma components in the predicted picture and the current picture;
filtering the residual value; and
attributing a metric to one of either the reference picture or the predicted picture based on the filtered residual value;
providing the predicted picture for display;
scaling the predicted picture according to the metric; and
combining the current picture with the scaled predicted picture.
2. The method of claim 1 , wherein the predicted picture is a motion predicted picture.
3. The method of claim 1 , wherein the picture to which the metric is attributed is the reference picture.
4. The method of claim 1 , wherein the picture to which the metric is attributed is the current picture.
5. The method of claim 1 , wherein the predicted picture is a motion compensated block, associated with a motion vector.
6. The method of claim 5 , wherein the method further comprises:
selecting the motion vector from a plurality of candidate motion vectors based on the metric.
7. The method of claim 5 , wherein an input video sequence comprises a previous video block followed by the current video picture, and wherein applying the motion vector to the previous video block generates the motion compensated block.
8. The method of claim 5 , wherein an input video sequence comprises the current video block followed by a future video block, and wherein applying the motion vector to the future video block generates the motion compensated block.
9. The method of claim 1 , wherein filtering is 2-dimensional.
10. The method of claim 9 , wherein filtering in the vertical dimension spans two pixels.
11. The method of claim 9 , wherein filtering in the vertical dimension spans three adjacent pixels, and wherein a center pixel is scaled by twice as much as a top pixel and a bottom pixel.
12. The method of claim 9 , wherein filtering in the vertical dimension spans five pixels.
13. The method of claim 9 , wherein filtering in the horizontal dimension spans a multiple of seven pixels.
14. A system for video evaluation in the presence of cross-chroma interference, said method comprising:
a motion compensator for generating a prediction block;
a residual generator for generating a residual block based on the prediction block and a current block;
a 2-dimensional filter for filtering the residual block; and
a metric generator for generating a metric associated with one of either the prediction block and the residual block based on the output from the 2-dimensional filter; and wherein the vertical dimension of the two-dimensional filters comprises three filter taps, and wherein a center tap has twice the weight as a top tap and a bottom tap.
15. The system of claim 14 , wherein the video block with which the metric is associated is the prediction block.
16. The system of claim 14 , wherein the video block with which the metric is associated is the current block.
17. The system of claim 14 , wherein the residual block comprises chroma differences between the residual block and the prediction block.
18. The system of claim 14 , wherein the vertical dimension of the two-dimensional filter comprises two filter taps of equal weight.
19. The system of claim 14 , wherein the vertical dimension of the two-dimensional filters comprises five filter taps of equal weight.
20. The system of claim 14 , wherein the horizontal dimension of the two-dimensional filters comprises seven filter taps of equal weight.
21. A method for video evaluation in the presence of cross-chroma interference, said method comprising:
generating a predicted picture based on a reference picture with a circuit;
generating a residual value based on a plurality of chroma components in the predicted picture and the current picture;
filtering the residual value;
attributing a metric to one of either the reference picture or the predicted picture based on the filtered residual value;
providing the predicted picture for display;
wherein the predicted picture is a motion compensated block, associated with a motion vector, selecting the motion vector from a plurality of candidate motion vectors based on the metric.
22. The method of claim 21 , wherein the predicted picture is a motion predicted picture.
23. The method of claim 21 , wherein the picture to which the metric is attributed is the reference picture.
24. The method of claim 21 , wherein the picture to which the metric is attributed is the current picture.
25. The method of claim 21 , wherein an input video sequence comprises a previous video block followed by the current video picture, and wherein applying the motion vector to the previous video block generates the motion compensated block.
26. The method of claim 21 , wherein an input video sequence comprises the current video block followed by a future video block, and wherein applying the motion vector to the future video block generates the motion compensated block.
27. The method of claim 21 , wherein filtering is 2-dimensional.
28. The method of claim 27 , wherein filtering in the vertical dimension spans two pixels.
29. The method of claim 27 , wherein filtering in the vertical dimension spans three adjacent pixels, and wherein a center pixel is scaled by twice as much as a top pixel and a bottom pixel.
30. The method of claim 27 , wherein filtering in the vertical dimension spans five pixels.
31. The method of claim 27 , wherein filtering in the horizontal dimension spans a multiple of seven pixels.
32. A method for video evaluation in the presence of cross-chroma interference, said method comprising:
generating a predicted picture based on a reference picture with a circuit;
generating a residual value based on a plurality of chroma components in the predicted picture and the current picture;
filtering the residual value;
attributing a metric to one of either the reference picture or the predicted picture based on the filtered residual value;
providing the predicted picture for display;
wherein the predicted picture is a motion compensated block; and
wherein an input video sequence comprises a previous video block followed by the current video picture, and wherein applying the motion vector to the previous video block generates the motion compensated block.
33. The method of claim 32 , wherein the predicted picture is a motion predicted picture.
34. The method of claim 32 , wherein the picture to which the metric is attributed is the reference picture.
35. The method of claim 32 , wherein the picture to which the metric is attributed is the current picture.
36. The method of claim 32 , wherein an input video sequence comprises the current video block followed by a future video block, and wherein applying the motion vector to the future video block generates the motion compensated block.
37. The method of claim 23 , wherein filtering is 2-dimensional.
38. The method of claim 37 , wherein filtering in the vertical dimension spans two pixels.
39. The method of claim 37 , wherein filtering in the vertical dimension spans three adjacent pixels, and wherein a center pixel is scaled by twice as much as a top pixel and a bottom pixel.
40. The method of claim 37 , wherein filtering in the vertical dimension spans five pixels.
41. The method of claim 37 , wherein filtering in the horizontal dimension spans a multiple of seven pixels.
42. A method for video evaluation in the presence of cross-chroma interference, said method comprising:
generating a predicted picture based on a reference picture with a circuit;
generating a residual value based on a plurality of chroma components in the predicted picture and the current picture;
filtering the residual value;
attributing a metric to one of either the reference picture or the predicted picture based on the filtered residual value;
providing the predicted picture for display;
wherein filtering is 2-dimensional; and
wherein filtering in the vertical dimension spans two pixels.
43. The method of claim 42 , wherein filtering in the vertical dimension spans three adjacent pixels, and wherein a center pixel is scaled by twice as much as a top pixel and a bottom pixel.
44. The method of claim 42 , wherein filtering in the vertical dimension spans five pixels.
45. The method of claim 42 , wherein filtering in the horizontal dimension spans a multiple of seven pixels.
46. A method for video evaluation in the presence of cross-chroma interference, said method comprising:
generating a predicted picture based on a reference picture with a circuit;
generating a residual value based on a plurality of chroma components in the predicted picture and the current picture;
filtering the residual value;
attributing a metric to one of either the reference picture or the predicted picture based on the filtered residual value;
providing the predicted picture for display;
wherein filtering is 2-dimensional; and wherein filtering in the vertical dimension spans three adjacent pixels, and wherein a center pixel is scaled by twice as much as a top pixel and a bottom pixel.
47. The method of claim 46 , wherein filtering in the vertical dimension spans five pixels.
48. The method of claim 46 , wherein filtering in the horizontal dimension spans a multiple of seven pixels.
49. A method for video evaluation in the presence of cross-chroma interference, said method comprising:
generating a predicted picture based on a reference picture with a circuit;
generating a residual value based on a plurality of chroma components in the predicted picture and the current picture;
filtering the residual value;
attributing a metric to one of either the reference picture or the predicted picture based on the filtered residual value;
providing the predicted picture for display;
wherein filtering is 2-dimensional; and
wherein filtering in the vertical dimension spans five pixels.
50. The method of claim 49 , wherein filtering in the horizontal dimension spans a multiple of seven pixels.
51. A method for video evaluation in the presence of cross-chroma interference, said method comprising:
generating a predicted picture based on a reference picture with a circuit;
generating a residual value based on a plurality of chroma components in the predicted picture and the current picture;
filtering the residual value;
attributing a metric to one of either the reference picture or the predicted picture based on the filtered residual value;
providing the predicted picture for display;
wherein filtering is 2-dimensional; and
wherein filtering in the horizontal dimension spans a multiple of seven pixels.
52. A system for video evaluation in the presence of cross-chroma interference, said method comprising:
a motion compensator for generating a prediction block;
a residual generator for generating a residual block based on the prediction block and a current block;
a 2-dimensional filter for filtering the residual block; and
a metric generator for generating a metric associated with one of either the prediction block and the residual block based on the output from the 2-dimensional filter; and wherein the vertical dimension of the two-dimensional filter comprises five filter taps of equal weight.
53. The system of claim 52 , wherein the video block with which the metric is associated is the prediction block.
54. The system of claim 52 , wherein the video block with which the metric is associated is the current block.
55. The system of claim 52 , wherein the residual block comprises chroma differences between the residual block and the prediction block.
56. The system of claim 52 , wherein the horizontal dimension of the two-dimensional filters comprises seven filter taps of equal weight.