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

1. A method of video coding using a coding mode belonging to a mode group comprising a regular Intra Prediction mode and a Matrix-based Intra Prediction (MIP) mode, the method comprising:

receiving input data related to a current block in a current picture at a video encoder side or receiving a video bitstream comprising coded data of the current block at a video decoder side;

when the MIP mode is applied to the current block:

checking availability for a target reference boundary region comprising left reference boundary samples and top reference boundary samples, wherein the left reference boundary samples and the top reference boundary samples are checked jointly for the current block;

deriving a reduced predictor, wherein said deriving the reduced predictor comprises applying matrix multiplication on a boundary vector and adding an offset, wherein the boundary vector is derived from the target reference boundary region, and wherein at least one unavailable samples is padded if there exists any unavailable sample in the target reference boundary region; and

encoding or decoding the current block using a MIP predictor derived from the reduced predictor.

2. The method of claim 1 , wherein the availability for the target reference boundary region is checked from a bottom sample to a top sample of the left reference boundary samples, then a top-left corner reference sample, and then from a leftmost sample to a rightmost sample of the top reference boundary samples.

3. The method of claim 1 , wherein if all reference samples in the target reference boundary region are not available, the target reference boundary region is padded with a default value.

4. The method of claim 1 , wherein if there exist at least one available sample and at least one unavailable sample in the target reference boundary region, any unavailable sample prior to a firstly available sample is padded using the firstly available sample.

5. The method of claim 4 , wherein a target unavailable sample after the firstly available sample is padded using a last available sample checked prior to the target unavailable sample.

6. The method of claim 1 , wherein the MIP predictor is derived by applying transpose, up-sampling or both to the reduced predictor.

7. An apparatus of video coding using a coding mode belonging to a mode group comprising a regular Intra Prediction mode and a Matrix-based Intra Prediction (MIP) mode, the apparatus of video coding comprising one or more electronic circuits or processors arranged to:

receive input data related to a current block in a current picture at a video encoder side or receive a video bitstream comprising coded data of the current block at a video decoder side;

when the MIP mode is applied to the current block:

check availability for a target reference boundary region comprising left reference boundary samples and top reference boundary samples, wherein the left reference boundary samples and the top reference boundary samples are checked jointly for the current block;

derive a reduced predictor, wherein a process to derive the reduced predictor comprises to apply matrix multiplication on a boundary vector and to add an offset, wherein the boundary vector is derived from the target reference boundary region, and wherein at least one unavailable samples is padded if there exists any unavailable sample in the target reference boundary region; and

encode or decode the current block using a MIP predictor derived from the reduced predictor.