Method and apparatus for updating predictions when coding motion information

Patent metadata

Field	Value
Publication number	US-9967589-B2
Application number	US-201113153289-A
Country	US
Kind code	B2
Filing date	Jun 3, 2011
Priority date	Jun 30, 2003
Publication date	May 8, 2018
Grant date	May 8, 2018

Abstract

Video compression algorithms typically represent visual information by a combination of motion and texture data. Motion data describes the temporal relationship between the content of a frame and that of a previous frame. This invention describes a method and apparatus for efficiently encoding motion data, particularly in the presence of horizontal and/or vertical motion boundaries.

First claim

I claim:

1. A method comprising:
setting, by a prediction generator, a first prediction motion vector and a second prediction motion vector, wherein the first prediction motion vector is derived from a left neighboring region of a block to which a motion vector to be decoded is associated, and wherein the second prediction motion vector is derived from an above neighboring region of the block to which the motion vector to be decoded is associated;
comparing, by a comparator, the first prediction motion vector and the second prediction motion vector; and
updating, by a controller, the second prediction motion vector using a third prediction motion vector when the first prediction motion vector is equal to the second prediction motion vector, an available third prediction motion vector exists in a video image, and a stop condition associated with a limit on a total number of predictions for at least one prediction generation process is not satisfied, wherein the first, second, and third prediction motion vectors are based on motion vectors at different positions,
wherein when the stop condition is satisfied regardless of whether additional prediction motion vectors exist in the video image or one or more of additional prediction motion vectors are not equal to the first prediction motion vector, the prediction generation process ends.

2. The method of claim 1 , further comprising determining a prediction motion vector candidate during coding of a motion vector.

3. The method of claim 2 , wherein the first prediction motion vector is equal to a motion vector associated with a first block located on a first side of a block associated with the motion vector being coded, the second prediction motion vector is equal to a motion vector associated with a block located on a second side of the block associated the motion vector being coded, and the third prediction motion vector is equal to a motion vector associated with a block located in proximity to the first block.

4. The method of claim 1 , further comprising:
determining whether the stop condition has been satisfied based on a predetermined number of prediction motion vectors that may be generated and compared with the first prediction motion vector, wherein the predetermined number comprises a limit on a total number of predictions.

5. The method of claim 1 , further comprising:
determining an updated prediction motion vector candidate based on a comparison result between the first prediction motion vector and third prediction motion vector.

6. A prediction generator comprising:
a memory for storing a first prediction motion vector and a second prediction motion vector, wherein the first prediction motion vector is derived from a left neighboring region of a block to which a motion vector to be decoded is associated, and wherein the second prediction motion vector is derived from an above neighboring region of the block to which the motion vector to be decoded is associated;
a comparator coupled with the memory to compare the first prediction motion vector and the second prediction motion vector; and
a controller coupled with the comparator to update the second prediction motion vector using a third prediction motion vector when the first prediction motion vector is equal to the second prediction motion vector, an available third prediction motion vector exists in a video image, and a stop condition associated with a limit on a total number of predictions for at least one prediction generation process is not satisfied, wherein the first, second, and third prediction motion vectors are based on motion vectors at different positions,
wherein when the stop condition is satisfied regardless of whether additional prediction motion vectors exist in the video image or one or more of additional prediction motion vectors are not equal to the first prediction motion vector, the prediction generation process ends.

7. The prediction generator of claim 6 , further comprising the controller to determine a prediction motion vector candidate during coding of a motion vector.

8. The prediction generator of claim 7 , wherein the first prediction motion vector is equal to a motion vector associated with a first block located on a first side of a block associated with the motion vector being coded, the second prediction motion vector is equal to a motion vector associated with a block located on a second side of the block associated the motion vector being coded, and the third prediction motion vector is equal to a motion vector associated with a block located in proximity to the first block.

9. The prediction generator of claim 6 , further comprising the controller to determine whether the stop condition has been satisfied based on a predetermined number of prediction motion vectors that may be generated and compared with the first prediction motion vector, wherein the predetermined number comprises a limit on a total number of predictions.

10. The prediction generator of claim 6 , further comprising the controller to determine an updated prediction motion vector candidate based on a comparison result between the first prediction motion vector and third prediction motion vector.

Assignees

• Bossen Frank Jan
• Ntt Docomo Inc

Inventors

• Bossen Frank Jan

Classifications

• H04N19/61

  in combination with predictive coding · CPC title

• H04N19/517 Primary

  by encoding · CPC title

• H04N19/52

  by predictive encoding · CPC title