Method and apparatus for intra prediction coding of video data involving matrix based intra-prediction

1. A method of decoding video data, comprising:

decoding, from a bitstream, a syntax element indicating an intra-prediction type of a current block of the video data, the intra-prediction type being indicated from among matrix based intra-prediction (MIP) and regular intra-prediction; and

generating a prediction block for the current block by selectively performing the MIP or the regular intra-prediction based on the intra-prediction type of the current block indicated by the syntax element,

wherein generating the prediction block for the current block by performing the MIP comprises:

decoding, from the bitstream, a syntax element indicating an MIP mode for the current block, the syntax element being represented as a truncated binary code specifying one of a plurality of MIP prediction modes allowed for a width and a height of the current block;

deriving an input boundary vector using neighboring samples adjacent to the current block based on the width and the height of the current block;

generating predicted samples for the current block based on matrix-vector multiplication between the input boundary vector and a matrix predefined for the MIP mode; and

deriving the prediction block for the current block based on the predicted samples,

wherein deriving the input boundary vector using the neighboring samples adjacent to the current block comprises:

generating an initial boundary vector filled with the neighboring samples adjacent to the current block or down-sampled values from the neighboring samples according to the width and the height of the current block; and

from the initial boundary vector, generating an input boundary vector to which the matrix-vector multiplication is applied, and

wherein generating the initial boundary vector comprises:

decoding, from the bitstream, a syntax element indicating a concatenation order of first entries of the initial boundary vector derived from the left neighboring samples adjacent to the current block and second entries of the initial boundary vector derived from the above neighboring samples adjacent to the current block; and

concatenating the first entries and the second entries according to the concatenation order and thereby generating the initial boundary vector.

2. The method of claim 1 , wherein deriving the prediction block for the current block based on the predicted samples comprises:

allocating the predicted samples to positions in the prediction block; and

generating predicted sample values for positions to which the predicted samples are not allocated in the prediction block, by performing horizontal interpolation and vertical interpolation on the predicted samples, left neighboring samples adjacent to the current block, and above neighboring samples adjacent to the current block.

3. The method of claim 2 , wherein the horizontal interpolation is performed prior to the vertical interpolation.

4. The method of claim 2 , wherein, before the horizontal interpolation and the vertical interpolation are performed, clipping is performed on the predicted samples such that the predicted samples lie between 0 and 2 bitDepth −1.

5. The method of claim 1 , wherein the current block is a luma block composed of a luma component,

wherein, when the MIP is performed on the luma block, and the intra-prediction mode of the luma block is used as an intra-prediction mode of a chroma block corresponding to the luma block, an intra-prediction mode of the chroma block is set as a PLANAR mode.

6. The method of claim 1 , wherein generating the prediction block for the current block by performing the regular intra-prediction comprises:

deriving Most Probable Mode (MPM) candidates based on a regular intra-prediction mode of each of neighboring blocks adjacent to the current block and thereby configuring an MPM list for the current block; and

deriving a regular intra-prediction mode for the current block based on the MPM list,

wherein, when an intra-prediction type of the neighboring blocks is the MIP, the regular intra-prediction mode of the neighboring block is regarded as a PLANAR mode.

7. A method of encoding video data, comprising:

encoding a syntax element indicating an intra-prediction type of a current block of the video data, the intra-prediction type being indicated from among matrix based intra-prediction (MIP) and regular intra-prediction;

generating a prediction block for the current block by selectively performing the MIP or the regular intra-prediction based on the intra-prediction type of the current block indicated by the syntax element; and

encoding a residual block that is a difference between the current block and the prediction block,

wherein generating the prediction block for the current block by performing the MIP comprises:

encoding a syntax element indicating an MIP mode for the current block, the syntax element being represented as a truncated binary code specifying one of a plurality of MIP prediction modes allowed for a width and a height of the current block;

deriving an input boundary vector using neighboring samples adjacent to the current block based on the width and the height of the current block;

generating predicted samples for the current block based on matrix-vector multiplication between the input boundary vector and a matrix predefined for the MIP mode; and

deriving the prediction block for the current block based on the predicted samples,

wherein deriving the input boundary vector using the neighboring samples adjacent to the current block comprises:

generating an initial boundary vector filled with the neighboring samples adjacent to the current block or down-sampled values from the neighboring samples according to the width and the height of the current block; and

from the initial boundary vector, generating an input boundary vector to which the matrix-vector multiplication is applied, and

wherein generating the initial boundary vector comprises:

encoding a syntax element indicating a concatenation order of first entries of the initial boundary vector derived from the left neighboring samples adjacent to the current block and second entries of the initial boundary vector derived from the above neighboring samples adjacent to the current block; and

concatenating the first entries and the second entries according to the concatenation order and thereby generating the initial boundary vector.

8. A non-transitory computer readable medium storing a bitstream containing encoded data for video data, the bitstream generated by processes of:

encoding a syntax element indicating an intra-prediction type of a current block of the video data, the intra-prediction type being indicated from among matrix based intra-prediction (MIP) and regular intra-prediction;

generating a prediction block for the current block by selectively performing the MIP or the regular intra-prediction based on the intra-prediction type of the current block indicated by the syntax element; and

encoding a residual block that is a difference between the current block and the prediction block,

wherein generating the prediction block for the current block by performing the MIP comprises:

encoding a syntax element indicating an MIP mode for the current block, the syntax element being represented as a truncated binary code specifying one of a plurality of MIP prediction modes allowed for a width and a height of the current block;

deriving an input boundary vector using neighboring samples adjacent to the current block based on the width and the height of the current block;

generating predicted samples for the current block based on matrix-vector multiplication between the input boundary vector and a matrix predefined for the MIP mode; and

deriving the prediction block for the current block based on the predicted samples,

wherein deriving the input boundary vector using the neighboring samples adjacent to the current block comprises:

generating an initial boundary vector filled with the neighboring samples adjacent to the current block or down-sampled values from the neighboring samples according to the width and the height of the current block; and

from the initial boundary vector, generating an input boundary vector to which the matrix-vector multiplication is applied, and

wherein generating the initial boundary vector comprises:

encoding a syntax element indicating a concatenation order of first entries of the initial boundary vector derived from the left neighboring samples adjacent to the current block and second entries of the initial boundary vector derived from the above neighboring samples adjacent to the current block; and

concatenating the first entries and the second entries according to the concatenation order and thereby generating the initial boundary vector.