Video compression for high efficiency video coding using a reduced resolution image
First Claim
Patent Images
1. A method of multi-level motion estimation for block-based video coding, the method comprising:
- constructing a reduced resolution image by;
representing 64×
64 blocks of original picture as 16×
16 blocks; and
calculating every pixel in the reduced image;
performing a reduced resolution search for blocks 64×
64;
performing refinement by neighboring blocks;
providing a motion assignment for 16×
16 blocks using neighboring blocks;
performing a refinement search for 16×
16 blocks;
providing an error distribution analysis information for split-merge prediction and non-square block selection;
performing a sub-pixel refinement;
splitting 16×
16 blocks into 8×
8 blocks;
merging four 16×
16 blocks into one 32×
32 partition;
merging four 32×
32 blocks into one 64×
64 partition; and
selecting a best non-square prediction using an error distribution field;
wherein the reduced resolution image is calculated as a weighted sum of pixels in the original picture, as follows;
Zx/4,y/4=(Ox,y+2*Ox+1,y+Ox+2,y+2*Ox,y+1+4*Ox+1,y+1+2*Ox+2,y+1+Ox,y+2+2*Ox+1,y+2+Ox+2,y+2)/16;
where;
Zx/4, y/4 is a pixel of the reduced image; and
Ox, y is a pixel in the original picture.
5 Assignments
0 Petitions
Accused Products
Abstract
Method for multi-level motion estimation in block-based video coding. The method is targeted to HEVC specifications of video compression, however, may be used with other video coding standards.
59 Citations
11 Claims
-
1. A method of multi-level motion estimation for block-based video coding, the method comprising:
-
constructing a reduced resolution image by; representing 64×
64 blocks of original picture as 16×
16 blocks; andcalculating every pixel in the reduced image; performing a reduced resolution search for blocks 64×
64;performing refinement by neighboring blocks; providing a motion assignment for 16×
16 blocks using neighboring blocks;performing a refinement search for 16×
16 blocks;providing an error distribution analysis information for split-merge prediction and non-square block selection; performing a sub-pixel refinement; splitting 16×
16 blocks into 8×
8 blocks;merging four 16×
16 blocks into one 32×
32 partition;merging four 32×
32 blocks into one 64×
64 partition; andselecting a best non-square prediction using an error distribution field;
wherein the reduced resolution image is calculated as a weighted sum of pixels in the original picture, as follows;
Zx/4,y/4=(Ox,y+2*Ox+1,y+Ox+2,y+2*Ox,y+1+4*Ox+1,y+1+2*Ox+2,y+1+Ox,y+2+2*Ox+1,y+2+Ox+2,y+2)/16;where; Zx/4, y/4 is a pixel of the reduced image; and Ox, y is a pixel in the original picture. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11)
-
Specification