Intra-prediction method and apparatus in video coding system

Patent metadata

Field	Value
Publication number	US-10750167-B2
Application number	US-201615768992-A
Country	US
Kind code	B2
Filing date	Sep 27, 2016
Priority date	Oct 22, 2015
Publication date	Aug 18, 2020
Grant date	Aug 18, 2020

Abstract

An image decoding method, which is performed by a decoding device according to the present invention, comprises the steps of: acquiring the information on an intra-prediction mode of a current block from a received bit stream; deriving neighboring reference samples of the current block; deriving a first prediction value for a current sample of the current block on the basis of the intra-prediction mode and the neighboring reference samples; deriving a second prediction value for the current sample on the basis of adjacent samples of the current sample; and deriving a prediction value for the current sample on the basis of the first prediction value and the second prediction value. According to the present invention, the intra-prediction value can be derived on the basis of the adjacent samples of the current sample within the present block, and the intra-prediction accuracy can be improved therethrough.

First claim

What is claimed is:

1. An image decoding method performed by a decoding device, comprising:
acquiring the information on an intra-prediction mode of a current block from a received bit stream;
deriving neighboring reference samples of the current block;
deriving a first prediction value for a current sample of the current block based on the intra-prediction mode and the neighboring reference samples;
deriving a second prediction value for the current sample based on adjacent samples of the current sample; and
deriving a prediction value for the current sample based on the first prediction value and the second prediction value,
wherein the image decoding method further comprising:
generating virtual reference samples for deriving the second prediction value,
at least one of the adjacent samples is included in the virtual reference samples, and
when a size of the current block is N×N and the coordinate of the upper left sample position of the current block is (0, 0),
the virtual reference samples prediction value are derived based on the following equation,

BR
=

(

A

2
⁢
N

-
1

+

L

2
⁢
N

-
1

)

/
2

B
x

=

L
N

+

diff
B

·

(

x
+
1

)

,

diff
B

=

BR
-

L
N

(

N
+
1

)

R
y

=

A
N

+

diff
R

·

(

y
+
1

)

,

diff
R

=

BR
-

A
N

(

N
+
1

)

where BR represents the virtual reference sample of coordinate (N, N), Bx represents the virtual reference sample of coordinate (x, N), Ry represents the virtual reference sample of coordinate (N, y), Ax represents a neighboring reference sample of coordinate (x, −1), and Ly represents the neighboring reference sample of coordinate (−1, y).

2. The decoding method of claim 1 , wherein the prediction value for the current sample is derived based on a weighted sum of the first prediction value and the second prediction value.

3. The decoding method of claim 2 , wherein the prediction value is derived based on the following equation,
α* P conv +(1−α)* P inner =P new

where P conv represents the first prediction value, P inner represents, the second prediction value, P new represents the prediction value, and a represents a weight.

4. The decoding method of claim 1 , wherein when a coordinate of an upper left sample position of the current block is (0, 0) and the coordinate of a sample position of the current sample is (x, y), the adjacent samples of the current sample includes a sample of coordinate (x−1, y−1), a sample of coordinate (x, y−1), and a sample of coordinate (x−1, y).

5. The decoding method of claim 4 , wherein the adjacent samples are located in the current block and the adjacent samples are reconstructed before the second prediction value for the current sample is derived.

6. The decoding method of claim 4 , wherein the second prediction value is derived based on the following equation,
P innerA ( x,y )=( A a *Rec ( x− 1, y− 1)+ A b *Rec ( x,y− 1)+ A c *Rec ( x− 1, y )+( A a +A b +A c )/2)/( A a +A b +A c )
where P innerA (x, y) represents the second prediction value for the current sample of the coordinate (x, y) in the current block, and Rec(x−1, y−1), Rec(x, y−1), and Rec(x−1, y) represent a reconstruction value of the sample of the coordinate (x−1, y−1), the reconstruction value of the sample of the coordinate (x, y−1), and the reconstruction value of the sample of the coordinate (x−1, y), respectively and A a , A b , and A c represent the weights for deriving the second prediction value.

7. The decoding method of claim 1 , wherein when the coordinate of the upper left sample position of the current block is (0, 0) and the coordinate of the sample position of the current sample is (x, y), the adjacent samples of the current sample include the sample of the coordinate (x−1, y), the sample of the coordinate (x−1, y+1), and the sample of the coordinate (x, y+1), and
the samples in the current block are sequentially reconstructed from a lower row in the current block to an upper row in the current block and sequentially reconstructed from a left sample to a right sample in each row.

8. The decoding method of claim 7 , wherein the second prediction value is derived based on the following equation,
P innerB ( x,y )=( B a *Rec ( x− 1, y )+ B b *Rec ( x− 1, y+ 1)+ B c *Rec ( x,y+ 1)+( B a +B b +B c )/2)/( B a +B b +B c )
where P innerB (x, y) re

Assignees

• Lg Electronics Inc
• Lg Electronics Inc

Inventors

• Son Eunyong
• Heo Jin
• Park Seungwook
• Yea Sehoon

Classifications

• H04N19/11

  among a plurality of spatial predictive coding modes · CPC title

• H04N19/593

  involving spatial prediction techniques · CPC title

• H04N19/159

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

• H04N19/176

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

• H04N19/52

  by predictive encoding · CPC title