METHOD AND APPARATUS FOR INTRABC MODE WITH FRACTIONAL-PEL BLOCK VECTOR RESOLUTION IN VIDEO CODING

• US 20180048909A1
• Filed: 03/02/2016
• Published: 02/15/2018
• Est. Priority Date: 03/02/2015
• Status: Active Grant

• Current Assignee: HFI Innovation, Inc. (MediaTek, Inc.)
• Original Assignee: HFI Innovation, Inc. (MediaTek, Inc.)
• Inventors: LIU, Shan, XU, Xiaozhong
• Granted Patent: US 10,362,331 B2
• Time in Patent Office: Days
• CPC Class Codes: H04N 19/105 Selection of the reference unit for prediction within a chosen coding or prediction mode, e.g. adaptive choice of position and number of pixels used for prediction
• CPC Class Codes: H04N 19/139 Analysis of motion vectors, e.g. their magnitude, direction, variance or reliability
• CPC Class Codes: H04N 19/159 Prediction type, e.g. intra-frame, inter-frame or bidirectional frame prediction
• CPC Class Codes: H04N 19/172 the region being a picture, frame or field
• CPC Class Codes: H04N 19/523 with sub-pixel accuracy
• CPC Class Codes: H04N 19/593 involving spatial prediction techniques

Abstract

A method and apparatus of video coding using IntraBC (Intra Block Copy) mode are disclosed. The system derives a block vector pointing from the current block to a reference block within a reconstructed pixel area of the current picture. If the block vector is at the fractional-pel resolution, the block vector is restricted so that neighboring pixels at integer-pel locations around fractional-pel reference pixels of the reference block for generating the fractional-pel reference pixels of the reference block using an interpolation filter are all available within the reconstructed pixel area. In another embodiment, the unavailable reference pixels for interpolation to generate fractional-pel reference pixels are padded.

Claims

1. A method of video encoding using IntraBC mode (Intra-block copy mode) coding for a current picture, comprising:

receiving a current block in the current picture;

deriving a block vector pointing from the current block to a reference block within a reconstructed pixel area of the current picture, wherein if fractional-pel resolution is enabled for the current block and the block vector is at the fractional-pel resolution, said deriving a block vector comprising restricting the block vector to cause neighboring pixels at integer-pel locations around fractional-pel reference pixels of the reference block for generating the fractional-pel reference pixels of the reference block to be all available within the reconstructed pixel area; and

encoding the current block according to IntraBC mode using the reference block as a predictor.

2. The method of claim 1, wherein the fractional-pel reference pixels of the reference block are generated from the neighboring pixels at integer-pel locations using an interpolation filter with 2 L filter tap length, wherein L is a positive integer.

3. The method of claim 2, wherein L is equal to 4 for a luma block and L is equal to 2 for a chroma block.

4. The method of claim 3, wherein said restricting the block vector shifts boundaries of the neighboring pixels at integer-pel locations available for generating the fractional-pel reference pixels of the reference block upward by at least (L−

1) pixels and leftward by at least (L−

1) pixels from boundaries of the reconstructed pixel area.

5. The method of claim 2, wherein whether the neighboring pixels at integer-pel locations around fractional-pel reference pixels of the reference block for generating the fractional-pel reference pixels of the reference block are all available within the reconstructed pixel area are determined according to if one or more conditions associated with the block vector and one or more coding parameters are satisfied.

6. The method of claim 5, wherein the current block corresponds to a PU (prediction unit), and said one or more coding parameters comprise PU width represented by nPbSw, PU height represented by nPbSh, location of a top-left pixel of the PU relative to the current picture represented by (xPb, yPb), location of a top-left pixel of a current CU (coding unit) relative to the current picture represented by (xCb, yCb), size of a CTU (coding tree unit) represented by CtbSize, and wherein the PU is within the current CU.

7. The method of claim 6, wherein said one or more conditions comprise

BV_x+nPbSw+xPb−

xCb+(L−

1)*2*(subWidthC−

1)<

=0 or

BV_y+nPbSh+yPb−

yCb+(L−

1)*2*(subHeightC−

1)<

=0.

8. The method of claim 7, wherein said one or more conditions further comprise

(xPb+BV_x+nPbSw−

1+(L−

1)*2*(subWidthC−

1))/CtbSize−

xCb/CtbSize yCb/CtbSize−

(yPb+BV_y+nPbSh−

1+(L−

1)*2*(subHeightC−

1))/CtbSize.

9. The method of claim 1, wherein the current block comprises a luma block and a chroma block for non-4:

4;

4 color video data, a luma block vector associated with the luma block is always at integer-pel resolution and a chroma block vector associated with the chroma block is scaled from the luma block vector, and said restricting the block vector is applied to the chroma block vector if the chroma block vector is at the fractional-pel resolution.

10. (canceled)

11. A method of video decoding using IntraBC mode (Intra-block copy mode) coding for a current picture, comprising:

receiving a video bitstream including coded data for a current block in the current picture;

parsing a current block vector from the video bitstream, wherein the current block vector points from the current block to a reference block within a reconstructed pixel area of the current picture;

if the current block vector is at a fractional-pel resolution, generating fractional-pel reference pixels for the reference block at fractional-pel locations using an interpolation filter, wherein if any filter input data used by the interpolation filter is unavailable, unavailable filter input data is padded for interpolation by the interpolation filter; and

decoding the current block according to IntraBC mode using the reference block as a predictor.

12. The method of claim 11, wherein the unavailable filter input data is padded by extending a rightmost available column of the reference block horizontally and extending a bottommost available row of the reference block vertically.

15. The method of claim 11, wherein the unavailable filter input data is padded using a default or pre-set value.

16. The method of claim 15, wherein the default or pre-set value corresponds to zero or ((1<

<

bit_depth)>

>

1), wherein bit_depth corresponds to pixel data bit depth of the current picture.

17. The method of claim 11, wherein the interpolation filter has 2 L filter tap length using L filter input data on each side of a filtered pixel, wherein L is a positive integer.

13. (canceled)

14. (canceled)

18. A method of video encoding using IntraBC mode (Intra-block copy mode) coding for a current picture, comprising:

receiving a current block in the current picture;

deriving a current block vector pointing from the current block to a reference block within a reconstructed pixel area of the current picture;

if the current block vector is at a fractional-pel resolution, generating fractional-pel reference pixels for the reference block at fractional-pel locations using an interpolation filter, wherein if any filter input data used by the interpolation filter is unavailable, unavailable filter input data is padded for interpolation by the interpolation filter; and

encoding the current block according to IntraBC mode using the reference block as a predictor.

19. The method of claim 18, wherein the unavailable filter input data is padded by extending a rightmost available column of the reference block horizontally and extending a bottommost available row of the reference block vertically.

22. The method of claim 18, wherein the unavailable filter input data is padded using a default or pre-set value.

23. The method of claim 22, wherein the default or pre-set value corresponds to zero or ((1<

<

bit_depth)>

>

1), wherein bit_depth corresponds to pixel data bit depth of the current picture.

24. The method of claim 18, wherein the interpolation filter has 2 L filter tap length using L filter input data on each side of a filtered pixel, wherein L is a positive integer.

20. (canceled)

21. (canceled)

25. An apparatus for video decoding using IntraBC mode (Intra-block copy mode) coding in a video decoding system, comprising one or more circuitries arranged to:

receive a video bitstream including coded data for a current block in a current picture;

parse a current block vector from the video bitstream, wherein the current block vector points from the current block to a reference block within a reconstructed pixel area of the current picture;

if the current block vector is at a fractional-pel resolution, generate fractional-pel reference pixels for the reference block at fractional-pel locations using an interpolation filter, wherein if any filter input data used by the interpolation filter is unavailable, unavailable filter input data is padded for interpolation by the interpolation filter; and

decode the current block according to IntraBC mode using the reference block as a predictor.