Composed prediction and restricted merge

Patent metadata

Field	Value
Publication number	US-12445595-B2
Application number	US-202418405290-A
Country	US
Kind code	B2
Filing date	Jan 5, 2024
Priority date	Feb 28, 2018
Publication date	Oct 14, 2025
Grant date	Oct 14, 2025

Abstract

A video decoder for decoding a video from a data stream using block-based predictive decoding is arranged to derive, using inter prediction, a first prediction, p 1 , for a current block, derive, using intra prediction, a second prediction, p 2 , for the current block determine a scalar weighting parameter, α, based at least in part on a look up table using prediction parameter information of blocks neighboring the current block, determine a prediction signal, q, for the current block as a weighted sum, of the first prediction, p 1 , and the second prediction, p 2 , using, at least in part, the scalar weighting parameter, α, and decode the current block using the prediction signal, q.

First claim

The invention claimed is:

1. A video decoder for decoding a video from a data stream using block-based predictive decoding, the video decoder comprising a processor configured to:
derive, using a first inter prediction mode, a first prediction, p 1 , for a current block;
set a first intermediate prediction signal, q 1 to the first prediction, p 1 ;
derive, using a first intra prediction mode, a second prediction, p 2 , for the current block;
derive, using a second prediction mode, a third prediction, p 3 , for the current block, the second prediction mode is either an intra prediction mode or an inter prediction mode;
determine multiple scalar weighting parameters, α n , based at least in part on a look up table using prediction parameter information of blocks neighboring the current block;
determine a second intermediate prediction signal, q 2 , for the current block as a weighted sum, of the first intermediate prediction signal, q 1 , and the second prediction, p 2 , using, at least in part, the scalar weighting parameter, α n ;
determine a composed prediction signal, q, for the current block as a weighted sum, of the second intermediate prediction signal, q 2 , and the third prediction, p 3 , using, at least in part, the scalar weighting parameter, α n ; and
decode the current block using the composed prediction signal, q.

2. The video decoder of claim 1 , wherein the weighted sum is equal to.

3. The video decoder of claim 1 , wherein the prediction parameter information of blocks which neighbor the current block is prediction modes of blocks which neighbor the current block.

4. A video encoder for encoding a video into a data stream using block-based predictive encoding, the video encoder comprising a processor configured to:
derive, using a first inter prediction mode, a first prediction, p 1 , for a current block;
set a first intermediate prediction signal, q 1 to the first prediction, p 1 ;
derive, using a first intra prediction mode, a second prediction, p 2 , for the current block;
derive, using a second prediction mode, a third prediction, p 3 , for the current block, the second prediction mode is either an intra prediction mode or an inter prediction mode;
determine multiple scalar weighting parameters, α n , based at least in part on a look up table using prediction parameter information of blocks neighboring the current block;
determine a second intermediate prediction signal, q 2 , for the current block as a weighted sum, of the first intermediate prediction signal, q 1 , and the second prediction, p 2 , using, at least in part, the scalar weighting parameter, α n ;
determine a composed prediction signal, q, for the current block as a weighted sum, of the second intermediate prediction signal, q 2 , and the third prediction, p 3 , using, at least in part, the scalar weighting parameter, α n ; and
encode the current block using the composed prediction signal, q.

5. The video encoder of claim 4 , wherein the weighted sum is equal to:
α n ·p n+1 +(1−α n )·q n .

6. The video encoder of claim 4 , wherein the prediction parameter information of blocks which neighbor the current block is prediction modes of blocks which neighbor the current block.

7. A method of decoding a video from a data stream using block-based predictive decoding, the method comprising:
deriving, using a first inter prediction mode, a first prediction, p 1 , for a current block;
set a first intermediate prediction signal, q 1 to the first prediction, p 1 ;
deriving, using a first intra prediction mode, a second prediction, p 2 , for the current block;
derive, using a second prediction mode, a third prediction, p 3 , for the current block, the second prediction mode is either an intra prediction mode or an inter prediction mode;
determining multiple scalar weighting parameters, α n , based at least in part on a look up table using prediction parameter information of blocks neighboring the current block;
determining a second intermediate prediction signal, q 2 , for the current block as a weighted sum, of the first intermediate prediction signal, q 1 , and the second prediction, p 2 , using, at least in part, the multiple scalar weighting parameters, α n ;
determining a composed prediction signal, q, for the current block as a weighted sum, of the second intermediate prediction signal, q 2 , and the third prediction, p 3 , using, at least in part, the multiple scalar weighting parameters, α n ; and
decoding the current block using the composed prediction signal, q.

8. The method of claim 7 , wherein the weighted sum is equal.

9. The method of claim 7 , wherein the prediction parameter information of blocks which neighbor the current block is prediction modes of blocks which neighbor the current block.

10. A method of encoding a video into a data stream using block-based predictive encoding, the method comprising:
deriving, using a first inter prediction mode, a first prediction, p 1 , for a current block;
set a first intermediate prediction signal, q 1 to the first prediction, p 1 ;
deriving, using a first intra prediction mode, a second prediction, p 2 , for the current block;
derive, using a second prediction mode, a third prediction, p 3 , for the current block, the second prediction mode is either an intra prediction mode or an inter prediction mode;
determining multiple scalar weighting parameters, α n , based at least in part on a look up table using prediction parameter information of blocks neighboring the current block;
determining a second intermediate prediction signal, q 2 , for the current block as a weighted sum, of the first intermediate prediction signal, q 1 , and the second prediction, p 2 , using, at least in part, the scalar weighting parameter, α n ;
determine a composed prediction signal, q, for the current block as a weighted sum, of the second intermediate prediction signal, q 2 , and the third prediction, p 3 , using, at least in part, the scalar weighting parameter, α n ; and
encoding the current block using the composed prediction signal, q.

11. The method of claim 10 , wherein the weighted sum is equal to.

12. The method of claim 10 , wherein the prediction parameter information of blocks which neighbor the current block is prediction modes of blocks which neighbor the current block.

13. A non-transitory digital storage medium having a computer program stored thereon to perform, when the computer program is run by a computer, a method of video decoding, the method comprising:
deriving, using a first inter prediction mode, a first prediction, p 1 , for a current block;
setting a first intermediate prediction signal, q 1 to the first prediction, p 1 ;
deriving, using a first intra prediction mode, a second prediction, p 2 , for the current block;
deriving, using a second prediction mode, a third prediction, p 3 , for the current block, the second prediction mode is either an intra prediction mode or an inter prediction mode;
determining multiple scalar weighting parameters, α n , based at least in part on a look up table using prediction parameter information of blocks neighboring the current block;
determining a second intermediate prediction signal, q 2 , for the current block as a weighted sum, of the first intermediate prediction signal, q 1 , and the second prediction, p 2 , using, at least in part, the scalar weighting parameter, α n ;
determining a composed prediction signal, q, for the current block as a weighted sum, of the second intermediate prediction signal, q 2 , and the third prediction, p 3 , using, at least in part, the scalar weighting parameter, α n ; and
decoding the current bl

Assignees

• Fraunhofer Ges Forschung

Inventors

• Wiegand Thomas
• Marpe Detlev
• Schwarz Heiko
• Winken Martin
• Bartnik Christian
• Pfaff Jonathan
• Helle Philipp
• Siekmann Mischa

Classifications

• H04N19/176

  the region being a block, e.g. a macroblock · CPC title

• H04N19/159

  Prediction type, e.g. intra-frame, inter-frame or bidirectional frame prediction · CPC title

• H04N19/11

  among a plurality of spatial predictive coding modes · CPC title

• H04N19/109

  among a plurality of temporal predictive coding modes · CPC title

• H04N19/107

  between spatial and temporal predictive coding, e.g. picture refresh · CPC title