IP Library Granted Patent US 12666027
Granted Patent B2
US 12666027 · App. 18/474,523 · Granted Jun 23, 2026

Systems and methods for blend intra mode coding

Inventors: Jing Ye (Palo Alto, CA); Xin Zhao (Palo Alto, CA); Liang Zhao (Palo Alto, CA); Han Gao (Palo Alto, CA); Shan Liu (Palo Alto, CA)
Assignee: TENCENT AMERICA LLC
H04N19/132H04N19/105H04N19/176
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12666027
App. No.
18/474,523
Granted
Jun 23, 2026
Kind
B2
Abstract

The various implementations described herein include methods and systems for encoding and decoding video. In one aspect, a method of video decoding includes receiving video data comprising a plurality of blocks, including a first block, from a video bitstream, where the first block is encoded in one of a plurality of smooth modes. The method also includes identifying a set of reference samples for the first block and deriving a first prediction value for the first block, where the first prediction value is derived from a second prediction value, the second prediction value is derived without a division or shift operation, and a final step in deriving the first prediction value comprises a division or shift operation. The method further includes decoding the first block using the first prediction value.

Claims (41)

1 . A method of video decoding performed at a computing system having memory and one or more processors, the method comprising:

receiving video data comprising a plurality of blocks, including a first block, from a video bitstream, wherein the first block is encoded in one of a plurality of smooth modes;

identifying a set of reference samples for the first block;

deriving a first prediction value for the first block, wherein the first prediction value is derived from a second prediction value, wherein the second prediction value is derived without a division or shift operation, and wherein a final step in deriving the first prediction value comprises a division or shift operation;

wherein, when the first block is encoded in a horizontal smooth mode, the second prediction value, P2, is derived according to P2=(L·(W−1−x)+TR·(x+1)·H, wherein W is a width of the first block, H is a height of the first block, L is a left reference sample from a location that is left of the first block, and TR is top-right reference sample from a location that is top right of the first block; and

wherein, when the first block is encoded in a vertical smooth mode, the second prediction value, P2, is derived according to P2=(T·(H−1−y)+BL·(y+1))·W, wherein T is a top reference sample from a location that is above the first block, and BL is bottom-left reference sample from a location that is bottom left of the first block; and

decoding the first block using the first prediction value.

2 . The method of claim 1 , further comprising deriving a fourth prediction value for the first block, wherein the first prediction value is derived from a combination of the second prediction value and the fourth prediction value.

3 . The method of claim 2 , wherein the fourth prediction value is derived without a division or shift operation.

4 . The method of claim 2 , wherein the fourth prediction value is a refined prediction value derived from a fifth prediction value.

5 . The method of claim 4 , wherein the fifth prediction value is derived via a second linear interpolation function involving a second reference sample of the set of reference samples and a height and a width of the first block.

6 . The method of claim 5 , wherein the second linear interpolation function does not include a division or shift operation.

7 . The method of claim 1 , wherein deriving the first prediction value for the first block includes performing a right shift operation.

8 . The method of claim 1 , wherein the first prediction value, P, is derived using P=(P′ H +P′ V +r)/(2·N·W·H), wherein P′ H is a horizontal prediction value, P′ V is a vertical prediction value, r is a rounding offset, and N is a predefined value.

9 . The method of claim 1 , wherein the first prediction value, P, is derived using P=(P′ H +P′ V +r)/(2·N·W·H/2 s V1 +s V2 ), wherein P′ H is a horizontal prediction value, P′ V is a vertical prediction value, r is a rounding offset, N is a predefined value, s V1 is a first shift value, and s V2 is a second shift value.

10 . The method of claim 1 , further comprising deriving a refined prediction value from the second prediction value, wherein the refined prediction value is derived without a division or shift operation, and wherein the first prediction value is derived from the refined prediction value.

11 . The method of claim 10 , wherein the refined prediction value is derived via a linear interpolation function involving a first reference sample of the set of reference samples and a height and a width of the first block.

12 . A method of video encoding performed at a computing system having memory and one or more processors, the method comprising:

receiving video data comprising a plurality of blocks, including a first block, wherein the first block is to be encoded in one of a plurality of smooth modes;

identifying a set of reference samples for the first block;

when the first block is to be encoded in a horizontal smooth mode, a first prediction value, P, is derived according to P=(L·(W−1−x)+TR·(x+1))·H, wherein W is a width of the first block, H is a height of the first block, L is a left reference sample from a location that is left of the first block, and TR is top-right reference sample from a location that is top right of the first block;

when the first block is to be encoded in a vertical smooth mode, the first prediction value, P, is derived according to P=(T·(H−1−y)+BL·(y+1))·W, wherein T is a top reference sample from a location that is above the first block, and BL is bottom-left reference sample from a location that is bottom left of the first block;

deriving a second prediction value for the first block, wherein the second prediction value is derived using the first prediction value, and wherein a final step in deriving the second prediction value comprises a division or shift operation; and

encoding the first block using the second prediction value.

13 . The method of claim 12 , further comprising deriving a refined prediction value from the first prediction value, wherein the refined prediction value is derived without a division or shift operation, and wherein the second prediction value is derived from the refined prediction value.

14 . The method of claim 13 , wherein the refined prediction value is derived via a linear interpolation function involving a first reference sample of the set of reference samples and a height and a width of the first block.

15 . The method of claim 12 , further comprising deriving a fourth prediction value for the first block, wherein the second prediction value is derived from a combination of the first prediction value and the fourth prediction value.

16 . The method of claim 12 , wherein deriving the second prediction value for the first block includes performing a right shift operation.

17 . A method of processing visual media data, the method comprising:

obtaining a source video sequence that comprises a plurality of frames; and

performing a conversion between the source video sequence and a video bitstream of visual media data according to a format rule;

wherein the video bitstream comprises a plurality of blocks, including a first block encoded in one of a plurality of smooth modes;

wherein the format rule specifies that:

a set of reference samples are to be identified for the first block;

when the first block is encoded in a horizontal smooth mode, a first prediction value, P, is to be derived according to P=(L·(W−1−x)+TR·(x+1))·H, wherein W is a width of the first block, H is a height of the first block, L is a left reference sample from a location that is left of the first block, and TR is top-right reference sample from a location that is top right of the first block;

when the first block is encoded in a vertical smooth mode, the first prediction value, P, is to be derived according to P=(T·(H−1−y)+BL·(y+1))·W, wherein T is a top reference sample from a location that is above the first block, and BL is bottom-left reference sample from a location that is bottom left of the first block;

a second prediction value is to be derived for the first block, the second prediction value being derived from the first prediction value, wherein a final step in deriving the second prediction value comprises a division or shift operation; and

the first block is to be decoded using the second prediction value.

18 . The method of claim 17 , wherein the format rule further specifies that a refined prediction value is to be derived from the first prediction value, wherein the refined prediction value is derived without a division or shift operation, and wherein the second prediction value is derived from the refined prediction value.

19 . The method of claim 18 , wherein the refined prediction value is derived via a linear interpolation function involving a first reference sample of the set of reference samples and a height and a width of the first block.

20 . The method of claim 17 , wherein the format rule further specifies that a fourth prediction value is to be derived for the first block, wherein the second prediction value is derived from a combination of the first prediction value and the fourth prediction value.