Methods for determining prediction value, encoder, and decoder

A method for determining a prediction value, being applicable to a decoder and comprising:

• parsing a bitstream to obtain a size of a current block and a coding mode of the current block;
• acquiring reconstructed values of neighboring samples of the current block and filtering the reconstructed values of the neighboring samples to obtain a reference value set of the current block, when the coding mode of the current block is a matrix-based intra prediction (MIP) mode;
• wherein in the reference value set, the number of values contained in filtered reference values of top neighboring samples and the number of values contained in filtered reference values of left neighboring samples are equal;
• calculating a second constant value according to a bit depth value of a luma component of a sample in the current block, when the size of the current block is smaller than a preset threshold value;
• assigning a difference between the second constant value and a first reference value in the reference value set as a first prediction input value in a prediction input value set;
• determining, according to the reference value set, other prediction input values in the prediction input value set other than the first prediction input value;
• calculating, according to the prediction input value set, prediction values of samples at specific positions in the current block;
• and
• performing interpolation filtering on the prediction values of the samples at the specific positions to obtain prediction values of samples at other positions in the current block other than the samples at the specific positions.

An encoder, comprising:

• a processor;
• and
• a storage medium storing instructions executable by the processor, wherein the storage medium is operable with the processor via a communication bus, and the instructions, when executed by the processor, are operable with the processor to:
• acquire reconstructed values of neighboring samples of a current block;
• filter the reconstructed values of the neighboring samples to obtain a reference value set of the current block;
• wherein in the reference value set, the number of values contained in filtered reference values of top neighboring samples and the number of values contained in filtered reference values of left neighboring samples are equal;
• calculate a first constant value according to a bit depth value of a luma component of a sample in the current block, when a size of the current block is smaller than a preset threshold value;
• determine a difference between the first constant value and a first reference value in the reference value set as a first prediction input value in a prediction input value set;
• determine, according to the reference value set, other prediction input values in the prediction input value set other than the first prediction input value;
• calculate, according to the prediction input value set, prediction values of samples at specific positions in the current block;
• and
• perform interpolation filtering on the prediction values of the samples at the specific positions to obtain prediction values of samples at other positions in the current block other than the samples at the specific positions.

A decoder, comprising:

• a processor;
• and
• a storage medium storing instructions executable by the processor, wherein the storage medium is operable with the processor via a communication bus, and the instructions, when executed by the processor, are operable with the processor to:
• parse a bitstream to obtain a size of a current block and a coding mode of the current block;
• acquire reconstructed values of neighboring samples of the current block and filter the reconstructed values of the neighboring samples to obtain a reference value set of the current block, when the coding mode of the current block is a matrix-based intra prediction (MIP) mode;
• wherein in the reference value set, the number of values contained in filtered reference values of top neighboring samples and the number of values contained in filtered reference values of left neighboring samples are equal;
• calculate a second constant value according to a bit depth value of a luma component of a sample in the current block, when the size of the current block is smaller than a preset threshold value;
• assign a difference between the second constant value and a first reference value in the reference value set as a first prediction input value in a prediction input value set;
• determine, according to the reference value set, other prediction input values in the prediction input value set other than the first prediction input value;
• calculate, according to the prediction input value set, prediction values of samples at specific positions in the current block;
• and
• perform interpolation filtering on the prediction values of the samples at the specific positions to obtain prediction values of samples at other positions in the current block other than the samples at the specific positions.