Method for coding image on basis of selective transform and device therefor
A method for decoding an image by a decoding device according to the present invention comprises the steps of: deriving, from a bit stream, transform coefficients of a block to be decoded; deriving residual samples of the block to be decoded, on the basis of a selective transform of the transform coefficients; and generating a reconstructed picture on the basis of the residual samples of the block to be decoded and prediction samples of the block to be decoded, wherein the selective transform is performed on the basis of a modified transform matrix, the modified transform matrix comprises a matrix including a modified basis vector, and the modified basis vector contains a certain number of elements selected from N elements.
1. An image decoding method performed by a decoding device, the method comprising:
deriving transform coefficients of a target block from a bitstream;
deriving residual samples for the target block based on a non-separable transform for the transform coefficients; and
generating a reconstructed picture based on the residual samples for the target block and prediction samples for the target block,
wherein the non-separable transform is performed based on a modified transform matrix,
wherein the modified transform matrix is a matrix including a modified basis vector,
wherein the modified basis vector includes fewer than N elements, and
wherein N is equal to the number of transform coefficients located in a region in which the non-separable transform is applied in the target block,
wherein the region in which the non-separable transform is applied is an 8×8 top-left target region in the target block and N is equal to 64.
2. The image decoding method of claim 1 ,
wherein the deriving of the residual samples for the target block based on the non-separable transform for the transform coefficients includes:
deriving modified transform coefficients by performing the non-separable transform on the transform coefficients located in the top-left target region of the target block based on the modified transform matrix and an association matrix including an association vector for the modified basis vector; and
deriving the residual samples by performing a core transform on the target block including the modified transform coefficients.
3. The image decoding method of claim 2 ,
wherein transform coefficients for elements with value 1 of the association vector are derived from among the transform coefficients located in the top-left target region, and a modified transform coefficient is derived based on the derived transform coefficients and the modified basis vector.
4. The image decoding method of claim 3 ,
wherein the association vector for the modified basis vector includes N number of elements, the N number of elements include the elements with value 1 and elements with value 0, the number of the elements with value 1 is A, and the modified basis vector includes A number of elements.
5. The image decoding method of claim 2 ,
wherein the modified transform matrix includes N number of modified basis vectors, and the association matrix includes N number of association vectors.
6. The image decoding method of claim 5 ,
wherein the association vectors include the same number of elements with value 1, and all the modified basis vectors include the same number of elements.
7. The image decoding method of claim 2 ,
wherein the modified transform matrix includes R number of modified basis vectors, and the association matrix includes R number of association vectors, and the R is less than N.
8. The image decoding method of claim 2 ,
wherein information on the association vector is obtained from a bitstream, and the association vector is derived based on the information on the association vector, and
wherein the information on the association vector includes information representing a starting point of elements with value 1, and information representing the number of elements with value 1.
9. An image encoding method performed by an encoding device, the method comprising:
deriving residual samples of a target block;
deriving transform coefficients of the target block by performing a non-separable transform based on the residual samples; and
encoding information on the transform coefficients,
wherein the non-separable transform is performed based on a modified transform matrix,
wherein the modified transform matrix is a matrix including a modified basis vector,
wherein the modified basis vector includes fewer than N elements, and
wherein N is equal to the number of transform coefficients located in a region in which the non-separable transform is applied in the target block,
wherein the region in which the non-separable transform is applied is an 8×8 top-left target region in the target block and N is equal to 64.
10. The image encoding method of claim 9 ,
wherein the deriving of the transform coefficients of the target block includes:
deriving modified transform coefficients by performing a core transform on the residual samples; and deriving the transform coefficients of the target block by performing the non-separable transform on modified transform coefficients located in the top-left target region of the target block based on the modified transform matrix and an association matrix including an association vector for the modified basis vector.
11. The image encoding method of claim 10 ,
wherein the modified transform matrix includes R number of modified basis vectors, and the association matrix includes R number of association vectors, and the R is less than N.
12. The image encoding method of claim 10 ,
wherein information on the association vector is entropy encoded, and the information on the association vector includes information representing a starting point of elements with value 1, and information representing the number of the elements with value 1.
13. A non-transitory computer-readable digital storage medium storing a bitstream generated by an image encoding method, the method comprising:
deriving residual samples of a target block;
deriving transform coefficients of the target block by performing a non-separable transform based on the residual samples; and
encoding information on the transform coefficients to generate the bitstream,
wherein the non-separable transform is performed based on a modified transform matrix,
wherein the modified transform matrix is a matrix including a modified basis vector,
wherein the modified basis vector includes fewer than N elements, and
wherein N is equal to the number of transform coefficients located in a region in
which the non-separable transform is applied in the target block,
wherein the region in which the non-separable transform is applied is a 8×8 top-left target region in the target block and N is equal to 64.
14. A transmission method for data comprising a bitstream for an image, the method comprising:
obtaining the bitstream for the image, wherein the bitstream is generated by
performing deriving residual samples of a target block, deriving transform coefficients of the target block by performing a non-separable transform based on the residual samples, and encoding information on the transform coefficients to generate the bitstream; and
transmitting the data comprising the bitstream,
wherein the non-separable transform is performed based on a modified transform matrix,
wherein the modified transform matrix is a matrix including a modified basis vector,
wherein the modified basis vector includes fewer than N elements, and
wherein N is equal to the number of transform coefficients located in a region in which the non-separable transform is applied in the target block,
wherein the region in which the non-separable transform is applied is a 8×8 top-left target region in the target block and N is equal to 64.