Image encoding/decoding method and device
The present invention relates to an image encoding/decoding method and apparatus. An image encoding method according to the present invention may comprise generating a transform block by performing at least one of transform and quantization; grouping at least one coefficient included in the transform block into at least one coefficient group (CG); scanning at least one coefficient included in the coefficient group; and encoding the at least one coefficient.
1. An image encoding method, comprising:
performing prediction for a current block using a prediction mode for the current block;
determining a transform kernel of the current block;
generating at least one coefficient of the current block by performing a transform on residual samples of the current block, based on the transform kernel; and
encoding transform kernel information indicating the transform kernel of the current block and the at least one coefficient of the current block,
wherein the transform comprises a primary transform and a secondary transform,
the at least one coefficient is generated by performing the secondary transform on a result of the primary transform on the residual samples using the transform kernel of the current block,
the transform kernel information indicates one of predefined transform kernel sets as the transform kernel of the current block,
each of the predefined transform kernel sets comprises a predefined horizontal transform kernel and a predefined vertical transform kernel,
the transform kernel of the current block comprises a horizontal transform kernel and a vertical transform kernel, and
the horizontal transform kernel and the vertical transform kernel of the transform kernel of the current block are determined based on a size of the current block, whether inter prediction is used for the prediction for the current block or not and whether intra prediction is used for the prediction for the current block or not.
2. The image encoding method of claim 1 , wherein the transform kernel information is index information indicating at least one of the predefined transform kernel sets.
3. The image encoding method of claim 1 , wherein the transform kernel information is encoded for each of coding units.
4. The image encoding method of claim 1 , wherein the transform kernel information is encoded when the size of the current block is less than or equal to a predefined size.
5. The image encoding method of claim 1 , wherein the transform kernel information is encoded when a non-zero transform coefficient exists in the current block.
6. The image encoding method of claim 1 , wherein the transform kernel information is encoded when a transform skip mode is not performed on the current block.
7. The image encoding method of claim 1 , wherein the horizontal transform kernel and the vertical transform kernel are determined independently.
8. An image decoding method, comprising:
decoding transform kernel information for a current block and at least one coefficient of the current block from a bitstream;
determining a transform kernel of the current block based on the transform kernel information; and
performing a transform on the at least one coefficient of the current block, based on the transform kernel to generate residual samples,
wherein prediction for the current block using a prediction mode for the current block is performed,
the transform comprises a primary transform and a secondary transform,
the residual samples are generated by performing the primary transform using the transform kernel of the current block on a result of the secondary transform on the at least one coefficient,
the transform kernel information indicates one of predefined transform kernel sets as the transform kernel of the current block,
each of the predefined transform kernel sets comprises a predefined horizontal transform kernel and a predefined vertical transform kernel,
the transform kernel of the current block comprises a horizontal transform kernel and a vertical transform kernel, and
the horizontal transform kernel and the vertical transform kernel of the transform kernel of the current block are determined based on a size of the current block, whether inter prediction is used for the prediction for the current block or not and whether intra prediction is used for the prediction for the current block or not.
9. The image decoding method of claim 8 , wherein the transform kernel information is index information indicating at least one of the predefined transform kernel sets.
10. The image decoding method of claim 8 , wherein the transform kernel information is decoded for each of coding units.
11. The image decoding method of claim 8 , wherein the transform kernel information is decoded when the size of the current block is less than or equal to a predefined size.
12. The image decoding method of claim 8 , wherein the transform kernel information is decoded when a non-zero transform coefficient exists in the current block.
13. The image decoding method of claim 8 , wherein the transform kernel information is decoded when a transform skip mode is not performed on the current block.
14. The image decoding method of claim 8 , wherein the horizontal transform kernel and the vertical transform kernel are determined independently.
15. A non-transitory computer-recordable medium for storing a bitstream, the bitstream comprising:
transform kernel information for a current block; and
at least one coefficient of the current block, wherein
a transform kernel of the current block is determined based on the transform kernel information,
a transform on the at least one coefficient of the current block is performed based on the determined transform kernel to generate residual samples,
prediction for the current block using a prediction mode for the current block is performed,
the transform comprises a primary transform and a secondary transform,
the residual samples are generated by performing the primary transform using the transform kernel of the current block on a result of the secondary transform on the at least one coefficient,
the transform kernel information indicates one of predefined transform kernel sets as the transform kernel of the current block,
each of the predefined transform kernel sets comprises a predefined horizontal transform kernel and a predefined vertical transform kernel,
the transform kernel of the current block comprises a horizontal transform kernel and a vertical transform kernel, and
the horizontal transform kernel and the vertical transform kernel of the transform kernel of the current block are determined based on a size of the current block, whether inter prediction is used for the prediction for the current block or not and whether intra prediction is used for the prediction for the current block or not.
16. The non-transitory computer-recordable medium of claim 15 , wherein the transform kernel information is index information indicating at least one of the predefined transform kernel sets.
17. The non-transitory computer-recordable medium of claim 15 , wherein the transform kernel information is encoded for each of coding units.
18. The non-transitory computer-recordable medium of claim 15 , wherein the transform kernel information is encoded when the size of the current block is less than or equal to a predefined size.
19. The non-transitory computer-recordable medium of claim 15 , wherein the transform kernel information is encoded when a non-zero transform coefficient exists in the current block.
20. The non-transitory computer-recordable medium of claim 15 , wherein the transform kernel information is encoded when a transform skip mode is not performed on the current block.