Image coding using transform index
An image decoding method according to the present document comprises the steps of: receiving image information including residual information for a current block; and generating residual samples of the current block on the basis of the residual information, wherein the residual information includes information on transform coefficients of the current block, and the residual samples are generated from the information on the transform coefficients on the basis of low-frequency non-separable transform (LFNST) index information.
1. An image decoding method performed by a decoding apparatus, the method comprising:
receiving image information including residual information for a current block;
deriving prediction samples of the current block based on intra prediction;
generating residual samples of the current block based on the residual information; and
generating reconstructed samples for the current block based on the prediction samples and the residual samples,
wherein the image information includes intra prediction type information for the current block,
wherein based on a tree type of the current block being single tree or dual tree luma, the intra prediction type information includes a matrix based intra prediction (MIP) flag representing whether MIP is applied to the current block,
wherein in response to the MIP being applied to the current block as the value of the MIP flag is equal to 1, the prediction samples of the current block are derived by reducing neighboring reference samples of the current block, performing a matrix-multiplication on the reduced neighboring reference samples, and adding an offset to values on which the matrix-multiplication is performed,
wherein a size of the matrix used for the matrix-multiplication is determined in response to a size of the current block,
wherein the size of the matrix is 16×8 in response to the size of the current block being 8×8,
wherein based on an intra sub-partitions (ISP) being not applied to the current block, a width and height of the current block being more than 4, an intra mode being applied to the current block, and the MIP flag representing that the MIP is not applied, the image information includes low frequency non-separable transform (LFNST) index information related to one of transform kernels in an LFNST transform set for the current block,
wherein a parsing order of the MIP flag always precedes a parsing order of the LFNST index information,
wherein based on the MIP flag representing that the MIP is applied, the image information does not include the LFNST index information,
wherein based on the LFNST index information being not present as the image information does not include the LFNST index information, a value of the LFNST index information is derived as 0,
wherein the residual information includes information on quantized transform coefficients of the current block, and
wherein the residual samples are generated based on the information on the quantized transform coefficients by using the LFNST index information.
2. An image encoding method performed by an encoding apparatus, the method comprising:
generating prediction samples of a current block by performing intra prediction on the current block;
generating intra prediction type information for the current block based on the performed intra prediction;
generating residual samples of the current block based on the prediction samples;
generating residual information including information on quantized transform coefficients for the current block based on the residual samples; and
encoding image information including the intra prediction type information and the residual information,
wherein based on a tree type of the current block being single tree or dual tree luma, the intra prediction type information includes a matrix based intra prediction (MIP) flag representing whether MIP is applied to the current block,
wherein in response to the value of the MIP flag equal to 1 as the MIP is applied to the current block, prediction samples of the current block are derived by reducing neighboring reference samples of the current block, performing a matrix-multiplication on the reduced neighboring reference samples, and adding an offset to values on which the matrix-multiplication is performed,
wherein a size of the matrix used for the matrix-multiplication is determined in response to a size of the current block,
wherein the size of the matrix is 16×8 in response to the size of the current block being 8×8,
wherein based on an intra sub-partitions (ISP) being not applied to the current block, a width and height of the current block being more than 4, an intra mode being applied to the current block, and the MIP flag representing that the MIP is not applied the image information includes low frequency non-separable transform (LFNST) index information related to one of transform kernels in an LFNST transform set for the current block,
wherein a signaling order of the MIP flag always precedes a signaling order of the LFNST index information,
wherein based on the MIP flag representing that the MIP is applied, the image information does not include the LFNST index information, and
wherein based on the LFNST index information being not present as the image information does not include the LFNST index information, a value of the LFNST index information is equal to 0.
3. A non-transitory computer-readable storage medium storing a bitstream generated by an encoding method, the method comprising:
generating prediction samples of a current block by performing intra prediction on the current block;
generating intra prediction type information for the current block based on the performed intra prediction;
generating residual samples of the current block based on the prediction samples;
generating residual information including information on quantized transform coefficients for the current block based on the residual samples; and
encoding image information including the intra prediction type information and the residual information,
wherein based on a tree type of the current block being single tree or dual tree luma, the intra prediction type information includes a matrix based intra prediction (MIP) flag representing whether MIP is applied to the current block,
wherein in response to the value of the MIP flag equal to 1 as the MIP is applied to the current block, prediction samples of the current block are derived by reducing neighboring reference samples of the current block, performing a matrix-multiplication on the reduced neighboring reference samples, and adding an offset to values on which the matrix-multiplication is performed,
wherein a size of the matrix used for the matrix-multiplication is determined in response to a size of the current block,
wherein the size of the matrix is 16×8 in response to the size of the current block being 8×8,
wherein based on an intra sub-partitions (ISP) being not applied to the current block, a width and height of the current block being more than 4, an intra mode being applied to the current block, and the MIP flag representing that the MIP is not applied the image information includes low frequency non-separable transform (LFNST) index information related to one of transform kernels in an LFNST transform set for the current block,
wherein a signaling order of the MIP flag always precedes a signaling order of the LFNST index information,
wherein based on the MIP flag representing that the MIP is applied, the image information does not include the LFNST index information, and
wherein based on the LFNST index information being not present as the image information does not include the LFNST index information, a value of the LFNST index information is equal to 0.
4. A transmission method of data for an image, the method comprising:
obtaining a bitstream for the image, the bitstream is generated by performing determining an intra prediction type for a current block, generating prediction samples of a current block by performing intra prediction on the current block, generating intra prediction type information for the current block based on the performed intra prediction, generating residual samples of the current block based on the prediction samples, generating residual information including information on quantized transform coefficients for the current block based on the residual samples, and encoding image information including the intra prediction type information and the residual information to generate the bitstream; and
transmitting the data comprising the bitstream,
wherein based on a tree type of the current block being single tree or dual tree luma, the intra prediction type information includes a matrix based intra prediction (MIP) flag representing whether MIP is applied to the current block,
wherein in response to the value of the MIP flag equal to 1 as the MIP is applied to the current block, prediction samples of the current block are derived by reducing neighboring reference samples of the current block, performing a matrix-multiplication on the reduced neighboring reference samples, and adding an offset to values on which the matrix-multiplication is performed,
wherein a size of the matrix used for the matrix-multiplication is determined in response to a size of the current block,
wherein the size of the matrix is 16×8 in response to the size of the current block being 8×8,
wherein based on an intra sub-partitions (ISP) being not applied to the current block, a width and height of the current block being more than 4, an intra mode being applied to the current block and the MIP flag representing that the MIP is not applied the image information includes low frequency non-separable transform (LFNST) index information related to one of transform kernels in an LFNST transform set for the current block,
wherein a signaling order of the MIP flag always precedes a signaling order of the LFNST index information,
wherein based on the MIP flag representing that the MIP is applied, the image information does not include the LFNST index information, and
wherein based on the LFNST index information being not present as the image information does not include the LFNST index information, a value of the LFNST index information is equal to 0.