Interpolation filter training method and apparatus, video picture encoding and decoding method, and encoder and decoder
Embodiments of this application disclose an interpolation filter training method and apparatus, a video picture encoding and decoding method, an encoder, and a decoder. According to the training method, a first sub-pixel picture obtained through interpolation by using a conventional interpolation filter is used as label data, to train a second interpolation filter, so that the second interpolation filter obtained through training can be directly used for a pixel value, obtained through interpolation, of a first fractional pixel position. Therefore, the label data is more accurate, and coding performance of a video picture is improved. According to the encoding method, during inter prediction, a target interpolation filter used for a current encoding picture block is determined from a set of candidate interpolation filters, and the encoder selects, according to content of the current encoding picture block, an appropriate interpolation filter to perform an interpolation operation.
1. A video picture encoding apparatus, comprising:
a non-transitory memory containing instructions; and
a processor in communication with the memory and upon execution of the instructions, is configured to:
perform an inter prediction process on a current encoding picture block to obtain motion information of the current encoding picture block, wherein the motion information of the current encoding picture block points to a fractional pixel position, and the inter prediction process comprises: determining, from a set of candidate interpolation filters based on a prediction block of the current encoding picture block being not an integer pixel picture block, a target interpolation filter used for the current encoding picture block; and
encode the current encoding picture block based on an inter prediction mode of the current encoding picture block and the motion information of the current encoding picture block to obtain encoding information, and encode the encoding information into a bitstream, wherein the encoding information comprises indication information of the target interpolation filter, and the indication information of the target interpolation filter is used to indicate to perform sub-pixel interpolation by using the target interpolation filter to obtain a reference block of a fractional pixel position corresponding to the current encoding picture block, wherein the set of candidate interpolation filters comprise an interpolation filter obtained by training data that is obtained through an interpolation of the fractional pixel position by using another interpolation filter of the set of candidate interpolation filters.
2. The apparatus according to claim 1 , wherein the processor is further configured to:
determine, from the set of candidate interpolation filters according to a rate-distortion cost criterion, the target interpolation filter used for the current encoding picture block.
3. The apparatus according to claim 1 , wherein the processor is further configured to:
determine an integer pixel reference picture block that optimally matches the current encoding picture block;
perform sub-pixel interpolation on the integer pixel reference picture block by using each interpolation filter in the set of candidate interpolation filters, to obtain N sub-pixel reference picture blocks, wherein N is a positive integer;
determine, from the integer pixel reference picture block and the N sub-pixel reference picture blocks, a prediction block that optimally matches the current encoding picture block; and
determine the motion information based on the prediction block, wherein an interpolation filter that is used to obtain the prediction block through interpolation is the target interpolation filter.
4. A video picture decoding apparatus, comprising:
a non-transitory memory containing instructions; and
a processor in communication with the memory and upon execution of the instructions, is configured to:
parse a bitstream to obtain indication information of a target interpolation filter, wherein the target interpolation filter is determined from a set of candidate interpolation filters based on a prediction block of a current decoding picture block being not an integer pixel picture block, and wherein the target interpolation filter is used for the current decoding picture block;
obtain motion information of the current decoding picture block, wherein the motion information points to a fractional pixel position;
perform a prediction process on the current decoding picture block based on the motion information of the current decoding picture block, wherein the prediction process comprises: performing sub-pixel interpolation based on the target interpolation filter indicated by the indication information, to obtain the prediction block of the current decoding picture block; and
reconstruct a reconstruction block of the current decoding picture block based on the prediction block of the current decoding picture block, wherein the target interpolation filter is an interpolation filter obtained by training data that is obtained through an interpolation of the fractional pixel position by using another interpolation filter.
5. The apparatus according to claim 4 , wherein the processor is further configured to:
parse the bitstream to obtain an index of the motion information of the current decoding picture block; and
determine the motion information of the current decoding picture block based on the index of the motion information of the current decoding picture block and a candidate motion information list of the current decoding picture block.
6. The apparatus according to claim 4 , wherein the processor is further configured to:
parse the bitstream to obtain an index of the motion information of the current decoding picture block and a motion vector difference;
determine a motion vector predictor of the current decoding picture block based on the index of the motion information of the current decoding picture block and a candidate motion information list of the current decoding picture block; and
obtain a motion vector of the current decoding picture block based on the motion vector predictor and the motion vector difference.
7. The apparatus according to claim 4 , wherein the processor is further configured to:
based on that an inter prediction mode of the current decoding picture block is a merge mode, obtain motion information that is of a previously decoded picture block and that is obtained through merging in the merge mode, that is, the motion information of the current decoding picture block.
8. A non-transitory computer-readable storage medium, comprising computer instructions that when executed by one or more processors, cause the one or more processors to perform the steps of:
parsing a bitstream to obtain indication information of a target interpolation filter, wherein the target interpolation filter is determined from a set of candidate interpolation filters based on a prediction block of a current decoding picture block being not an integer pixel picture block, and wherein the target interpolation filter is used for the current decoding picture block;
obtaining motion information of the current decoding picture block, wherein the motion information points to a fractional pixel position;
performing a prediction process on the current decoding picture block based on the motion information of the current decoding picture block, wherein the prediction process comprises: performing sub-pixel interpolation based on the target interpolation filter indicated by the indication information, to obtain the prediction block of the current decoding picture block; and
reconstructing a reconstruction block of the current decoding picture block based on the prediction block of the current decoding picture block, wherein the target interpolation filter is an interpolation filter obtained by training data that is obtained through an interpolation of the fractional pixel position by using another interpolation filter.
9. The medium according to claim 8 , wherein the obtaining the motion information of the current decoding picture block comprises:
parsing the bitstream to obtain an index of the motion information of the current decoding picture block; and
determining the motion information of the current decoding picture block based on the index of the motion information of the current decoding picture block and a candidate motion information list of the current decoding picture block.
10. The medium according to claim 8 , wherein the obtaining the motion information of the current decoding picture block comprises:
parsing the bitstream to obtain an index of the motion information of the current decoding picture block and a motion vector difference;
determining a motion vector predictor of the current decoding picture block based on the index of the motion information of the current decoding picture block and a candidate motion information list of the current decoding picture block; and
obtaining a motion vector of the current decoding picture block based on the motion vector predictor and the motion vector difference.
11. The medium according to claim 8 , wherein the obtaining the motion information of the current decoding picture block comprises:
based on that an inter prediction mode of the current decoding picture block is a merge mode, obtaining motion information that is of a previously decoded picture block and that is obtained through merging in the merge mode, that is, the motion information of the current decoding picture block.