IP Library Granted Patent US 10,819,987
Granted Patent B2
US 10,819,987 · App. 15/775,307 · Granted Oct 27, 2020

Method and apparatus for coefficient induced intra prediction in image coding system

Inventors: Hyeongmoon Jang (Seoul, KR); Bumshik Lee (Seoul, KR); Sehoon Yea (Seoul, KR)
Assignee: LG ELECTRONICS INC.
H04N19/18H04N19/105H04N19/11H04N19/119H04N19/124H04N19/159H04N19/176H04N19/44H04N19/45H04N19/50H04N19/61
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 10,819,987
App. No.
15/775,307
Granted
Oct 27, 2020
Kind
B2
Abstract

An intra prediction method according to the present invention comprises the steps of: determining an intra prediction mode for a current block; determining whether to apply a coefficient induced intra prediction (CIIP) to the current block; deriving neighboring reference samples for the current block on the basis of whether the CIIP is applied; and generating prediction samples for the current block on the basis of the intra prediction mode and the neighboring samples. The present invention can generate modified or improved neighboring reference samples and thereby can improve the performance of intra prediction.

Claims (54)

1. An intra prediction method performed by a decoding device, comprising:

determining an intra prediction mode for a current block;

determining whether to apply a coefficient induced intra prediction (CIIP) to the current block;

deriving neighboring reference samples for the current block on the basis of whether to apply the CIIP; and

generating prediction samples for the current block on the basis of the intra prediction mode and the neighboring reference samples,

wherein the deriving of the neighboring reference samples comprises:

acquiring quantized transform coefficients for a neighboring block;

acquiring transform coefficients by inversely quantizing the quantized transform coefficients;

deriving residual samples or modified residual samples on the basis of the transform coefficients; and

deriving the neighboring reference samples on the basis of the residual samples or the modified residual samples,

wherein if the CIIP is not applied to the current block, the residual samples are derived by inversely transforming the transform coefficients, and

wherein if the CIIP is applied to the current block, the modified residual samples are derived by inversely transforming transform coefficients subjected to CIIP scaling after a CIIP scaling procedure is performed on the transform coefficients,

wherein the CIIP scaling procedure is performed based on a scaling matrix, and

wherein the scaling matrix is determined according to a intra prediction mode for the neighboring block and a transform kernel, and

wherein the transform coefficients subjected to the CIIP scaling are inversely transformed based on the transform kernel.

2. The intra prediction method of claim 1 , further comprising receiving at least one of information on a CIIP available flag and information on a CIIP available block size, wherein whether to apply the CIIP is determined on the basis of at least one of the information on the CIIP available flag and the information on the CIIP available block size.

3. The intra prediction method of claim 2 , wherein the information on the CIIP available block size is received when a value of the CIIP available flag is 1.

4. The intra prediction method of claim 2 ,

wherein the information on the CIIP available block size indicates a maximum block size and a minimum block size for which the CIIP is available, and

wherein if a size of the current block belongs to a range of the maximum block size and the minimum block size, it is determined that the CIIP is applied to the current block.

5. The intra prediction method of claim 1 ,

wherein the current block corresponds to a transform unit (TU),

wherein the TU is located in a current prediction unit (PU), and

wherein if the current PU is a non-square PU, it is determined that the CIIP is not applied to the current block.

6. The intra prediction method of claim 1 ,

wherein neighboring reference samples for the current block are derived based on modified reconstructed samples for a first neighboring block and reconstructed samples for a second neighboring block,

wherein the deriving of the neighboring reference samples comprises:

acquiring first quantized transform coefficients for the first neighboring block;

acquiring second quantized transform coefficients for the second neighboring block;

acquiring first transform coefficients by inversely quantizing the first quantized transform coefficients;

acquiring second transform coefficients by inversely quantizing the second quantized transform coefficients;

deriving transform coefficients subjected to CIIP scaling by performing a CIIP scaling procedure on the first transform coefficients;

deriving modified residual samples by inversely transforming the transform coefficients subjected to CIIP scaling;

deriving residual samples by inversely transforming the second transform coefficients;

deriving the modified reconstructed samples for the first neighboring block on the basis of the modified residual samples; and

deriving the reconstructed samples for the second neighboring block on the basis of the residual samples.

7. The intra prediction method of claim 6 , wherein if there is a non-available neighboring reference sample among the neighboring reference samples for the current block, a previous available neighboring reference sample is copied based on a predetermined search order to perform padding on the non-available neighboring reference sample.

8. A decoding device for performing intra prediction, comprising:

a predictor for determining an intra prediction mode for a current block, determining whether to apply a coefficient induced intra prediction (CIIP) to the current block, deriving neighboring reference samples for the current block on the basis of whether to apply the CIIP, and generating prediction samples for the current block on the basis of the intra prediction mode and the neighboring reference samples; and

a memory for storing coding information for at least one neighboring block,

wherein the deriving of the neighboring reference samples comprises:

acquiring quantized transform coefficients for a neighboring block;

acquiring transform coefficients by inversely quantizing the quantized transform coefficients;

deriving residual samples or modified residual samples on the basis of the transform coefficients; and

deriving the neighboring reference samples on the basis of the residual samples or the modified residual samples,

wherein if the CIIP is not applied to the current block, the residual samples are derived by inversely transforming the transform coefficients, and

wherein if the CIIP is applied to the current block, the modified residual samples are derived by inversely transforming transform coefficients subjected to CIIP scaling after a CIIP scaling procedure is performed on the transform coefficients,

wherein the CIIP scaling procedure is performed based on a scaling matrix, and

wherein the scaling matrix is determined according to a intra prediction mode for the neighboring block and a transform kernel, and

wherein the transform coefficients subjected to the CIIP scaling are inversely transformed based on the transform kernel.

9. The decoding device of claim 8 , further comprising an entropy decoder for receiving at least one of information on a CIIP available flag and information on a CIIP available block size, wherein the predictor determines whether to apply the CIIP on the basis of at least one of the information on the CIIP available flag and the information on the CIIP available block size.

10. The decoding device of claim 9 ,

wherein the information on the CIIP available block size indicates a maximum block size and a minimum block size for which the CIIP is available, and

wherein the predictor determines that the CIIP is applied to the current block if a size of the current block belongs to a range of the maximum block size and the minimum block size.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Aug 28, 2018
From: JANG, HYEONGMOON; LEE, BUMSHIK
To: LG ELECTRONICS INC.
Reel/Frame 046728/0126 →
Continuity (2)
Provisional Application 62254669 · Nov 12, 2015
Related Publication 20180343455A1 · Nov 29, 2018
Cited By (7)
US 12,273,527 US 12,348,713 US 12,382,105 US 12,425,585 US 12,470,729 US 12,526,407 US 12,621,445