Template-based inter prediction techniques based on encoding and decoding latency reduction
Video coding methods are described for reducing latency in template-based inter coding. In some embodiments, a method is provided for coding a video that includes a current picture and at least one reference picture. For at least a current block in the current picture, a respective predicted value is generated (e.g. using motion compensated prediction) for each sample in a template region adjacent to the current block. Once the predicted values are generated for each sample in the template region, a process is invoked to determine a template-based inter prediction parameter by using predicted values in the template region and sample values the reference picture. This process can be invoked without waiting for reconstructed sample values in the template region. Template-based inter prediction of the current block is then performed using the determined template-based inter prediction parameter.
1. A method for video encoding, comprising:
dividing a current picture into a plurality of separate template slices, each template slice comprising a plurality of blocks;
signaling at least one of a number of template slices in said current picture and a position of each template slice in said current picture at a sequence level or a picture level;
determining an inter prediction mode for a block in a current template slice, wherein the inter prediction mode is selected from among at least one template-based inter prediction mode and at least one non-template-based inter prediction mode; and
generating a prediction for each block in the current template slice, wherein the prediction of a block in the current template slice using a template-based inter prediction mode is based on samples in a template region for the block and is constrained from using for the prediction any samples that are in the current picture but are outside the current template slice, wherein non-template-based inter prediction modes are allowed to use, for the prediction, samples that are outside the current template slice, and wherein the template region includes previously reconstructed neighboring samples, in the current template slice, for the block.
2. The method of claim 1 , wherein in-loop filter and intra prediction are allowed to operate across template slice boundaries.
3. The method of claim 1 , wherein blocks in different template slices in the current picture are encoded in parallel.
4. A non-transitory machine readable medium having stored thereon machine executable instructions that, when executed, implement a method of video encoding according to claim 1 .
5. An apparatus for video encoding, comprising at least a memory and one or more processors, wherein said one or more processors are configured to:
divide a current picture into a plurality of template slices, each template slice comprising a plurality of blocks;
signal at least one of a number of template slices in said current picture and a position of each template slice in said current picture at a sequence level or a picture level;
determine an inter prediction mode for each block in a current template slice, wherein the inter prediction mode is selected from among at least one template-based inter prediction mode and at least one non-template-based inter prediction mode; and
generate a prediction for each block in the current template slice, wherein the prediction of a block in the current template slice using a template-based inter prediction mode is based on samples in a template region for the block and is constrained from using for the prediction any samples that are in the current picture but are outside the current template slice, wherein non-template-based inter prediction modes are allowed to use, for the prediction, samples that are outside the current template slice, and wherein the template region includes previously reconstructed neighboring samples, in the current template slice, for the block.
6. The apparatus of claim 5 , wherein in-loop filters and intra prediction are allowed to operate across template slice boundaries.
7. The apparatus of claim 5 , wherein non-template-based inter prediction modes are not constrained from using, for the prediction, coding information from blocks that are outside the current template slice.
8. The apparatus of claim 5 , wherein a number of coding tree units (CTUs) in each template slice is signaled in a bitstream.
9. The apparatus of claim 5 , wherein blocks in different template slices in the current picture are decoded in parallel.
10. A method for video decoding, comprising:
dividing a current picture into a plurality of template slices, each template slice comprising a plurality of blocks;
decoding at least one of a number of template slices in said current picture and a position of each template slice in said current picture signaled at a sequence level or a picture level;
determining an inter prediction mode for a block in a current template slice, wherein the inter prediction mode is selected from among at least one template-based inter prediction mode and at least one non-template-based inter prediction mode; and
generating a prediction for each block in the current template slice, wherein the prediction of a block in the current template slice using a template-based inter prediction mode is based on samples in a template region for the block and is constrained from using for the prediction any samples that are in the current picture but are outside the current template slice, wherein non-template-based inter prediction modes are allowed to use, for the prediction, samples that are outside the current template slice, and wherein the template region includes previously decoded neighboring samples, in the current template slice, for the block.
11. The method of claim 10 , wherein in-loop filters and intra prediction are allowed to operate across template slice boundaries.
12. The method of claim 10 , wherein blocks in different template slices in the current picture are decoded in parallel.
13. A non-transitory machine readable medium having stored thereon machine executable instructions that, when executed, implement a method of video decoding according to claim 10 .
14. An apparatus for video decoding, comprising at least a memory and one or more processors, wherein said one or more processors are configured to:
divide a current picture into a plurality of template slices, each template slice comprising a plurality of blocks;
decode at least one of a number of template slices in said current picture and a position of each template slice in said current picture signaled at a sequence level or a picture level;
determine an inter prediction mode for each block in a current template slice, wherein the inter prediction mode is selected from among at least one template-based inter prediction mode and at least one non-template-based inter prediction mode; and
generate a prediction for each block in the current template slice, wherein the prediction of a block in the current template slice using a template-based inter prediction mode is based on samples in a template region for the block and is constrained from using for the prediction any samples that are in the current picture but are outside the current template slice, wherein non-template-based inter prediction modes are allowed to use, for the prediction, samples that are outside the current template slice, and wherein the template region includes previously decoded neighboring samples, in the current template slice, for the block.
15. The apparatus of claim 14 , wherein in-loop filters and intra prediction are allowed to operate across template slice boundaries.
16. The apparatus of claim 14 , wherein non-template-based inter prediction modes are not constrained from using, for the prediction, coding information from blocks that are outside the current template slice.
17. The apparatus of claim 14 , wherein a number of coding tree units (CTUs) in each template slice is signaled in a bitstream.
18. The apparatus of claim 14 , wherein blocks in different template slices in the current picture are decoded in parallel.