Context derivation for arithmetic coding of transform coefficients generated by non-separable transforms
A video bitstream including a current transform block (TB) in a current picture is received. A context model is determined for a syntax element associated with a transform coefficient level of a first coefficient group (CG) in the current TB based on transform coefficient levels of at least one first neighboring CG of the first CG. The first CG is positioned on a first scanning line. The at least one first neighboring CG is positioned on a second scanning line that is scanned before the first scanning line. The context model is a probability model for a non-separable transform. The first CG is reconstructed based on the transform coefficient level that is determined according to the determined the context model.
1 . A method of video decoding, the method comprising:
receiving a video bitstream including a current transform block (TB) in a current picture;
determining a context model for a syntax element associated with a transform coefficient level of a first coefficient group (CG) in the current TB based on transform coefficient levels of at least one first neighboring CG of the first CG, the first CG being positioned on a first scanning line, the at least one first neighboring CG being positioned on a second scanning line that is scanned before the first scanning line, the context model being a probability model for a non-separable transform, each of a plurality of CGs in the first scanning line being coded according to a prediction mode based on a different group of closest neighboring CGs according to a position of the respective CG in the first scanning line; and
reconstructing the first CG based on the transform coefficient level that is determined according to the determined the context model.
2 . The method of claim 1 , wherein the at least one first neighboring CG includes a first neighboring CG positioned on the second scanning line and another first neighboring CG positioned on a third scanning line that is scanned before the first scanning line and the second scanning line.
3 . The method of claim 1 , wherein the at least one first neighboring CG is positioned further than the first CG from a top-left sample position of the current TB.
4 . The method of claim 1 , wherein the first scanning line and the second scanning line are parallel diagonal lines.
5 . The method of claim 1 , wherein a number of the at least one first neighboring CG is determined based on a position of the first CG in the current TB.
6 . The method of claim 1 , wherein a number of the at least one first neighboring CG is in a range from 1 to 15.
7 . The method of claim 1 , wherein the syntax element is associated with an absolute value of the transform coefficient level of the first CG.
8 . The method of claim 1 , wherein the non-separable transform is one of a low-frequency non-separable transform (LFNST) mode and a non-separable primary transforms (NSPT) mode.
9 . The method of claim 1 , further comprising:
determining a context model for a syntax element associated with a transform coefficient level for each of a plurality of CGs positioned on the first scanning line based on the transform coefficient levels of the at least one first neighboring CG of the first CG.
10 . The method of claim 1 , further comprising:
determining a context model for a syntax element associated with a transform coefficient level of a second CG on the first scanning line based on transform coefficient levels of at least one second neighboring CG of the second CG, the at least one second neighboring CG including a second neighboring CG positioned on the second scanning line and another second neighboring CG that is different from the at least one first neighboring CG.
11 . A method of video encoding, the method comprising:
determining a context model for a syntax element associated with a transform coefficient level of a first coefficient group (CG) in a current transform block (TB) of a current picture based on transform coefficient levels of at least one first neighboring CG of the first CG, the first CG being positioned on a first scanning line, the at least one first neighboring CG being positioned on a second scanning line that is scanned before the first scanning line, the context model being a probability model for a non-separable transform, each of a plurality of CGs in the first scanning line being coded according to a prediction mode based on a different group of closest neighboring CGs according to a position of the respective CG in the first scanning line;
encoding the first CG in a bitstream based on the transform coefficient level that is determined according to the determined the context model; and
transmitting the encoded bitstream.
12 . The method of claim 11 , wherein the at least one first neighboring CG includes a first neighboring CG positioned on the second scanning line and another first neighboring CG positioned on a third scanning line that is scanned before the first scanning line and the second scanning line.
13 . The method of claim 11 , wherein the at least one first neighboring CG is positioned further than the first CG from a top-left sample position of the current TB.
14 . The method of claim 11 , wherein the first scanning line and the second scanning line are parallel diagonal lines.
15 . The method of claim 11 , wherein a number of the at least one first neighboring CG is determined based on a position of the first CG in the current TB.
16 . The method of claim 11 , wherein a number of the at least one first neighboring CG is in a range from 1 to 15.
17 . The method of claim 11 , wherein the syntax element is associated with an absolute value of the transform coefficient level of the first CG.
18 . The method of claim 11 , wherein the non-separable transform is one of a low-frequency non-separable transform (LFNST) mode and a non-separable primary transforms (NSPT) mode.
19 . The method of claim 11 , further comprising:
determining a context model for a syntax element associated with a transform coefficient level for each of a plurality of CGs positioned on the first scanning line based on the transform coefficient levels of the at least one first neighboring CG of the first CG.
20 . A non-transitory computer-readable storage medium storing instructions which when executed by a processor cause the processor to perform an encoding method comprising:
determining a context model for a syntax element associated with a transform coefficient level of a first coefficient group (CG) in a current transform block (TB) of a current picture based on transform coefficient levels of at least one first neighboring CG of the first CG, the first CG being positioned on a first scanning line, the at least one first neighboring CG being positioned on a second scanning line that is scanned before the first scanning line, the context model being a probability model for a non-separable transform, each of a plurality of CGs in the first scanning line being coded according to a prediction mode based on a different group of closest neighboring CGs according to a position of the respective CG in the first scanning line;
encoding the first CG in a bitstream based on the transform coefficient level that is determined according to the determined the context model; and
transmitting the encoded bitstream.