Unified neural network filter model
A method implemented by a video coding apparatus. The method includes applying a first filter to an unfiltered sample of a video unit to generate a filtered sample. The first filter is a neural network (NN) filter based on a non-deep learning-based filter (NDLF) being disabled, and the first filter is the NDLF based on the NN filter being disabled. The method also includes performing a conversion between a video media file and a bitstream based on the filtered sample that was generated.
1 . A method for processing video data, comprising:
determining, for a conversion between a video unit of a video and a bitstream of the video, interaction between a first filtering tool and a second filtering tool for being applied to the video unit, wherein the first filtering tool comprises a neural network (NN) filter, and the second filtering tool comprises a non-deep learning-based filtering (NDLF) tool; and
performing the conversion based on the determining,
wherein the interaction between the first filtering tool and the second filtering tool depends on a color format or a color component of the video unit, or
wherein the interaction between the first filtering tool and the second filtering tool depends on a profile flag, a tier flag, a level flag, or a constraint flag.
2 . The method of claim 1 , wherein when the NN filter is applied to the video unit, one or more filters of the NDLF tool are disabled for the video unit.
3 . The method of claim 2 , wherein the NDLF tool comprises a deblocking (DB) filter, a sample adaptive offset (SAO) filter, an adaptive loop filter (ALF), a cross-component (CC)-SAO filter, a CC-ALF, a luma mapping with chroma scaling (LMCS) filter, a bilateral filter, or a transform-domain filter.
4 . The method of claim 3 , wherein the NN filter is applied to the video unit based on the ALF being disabled.
5 . The method of claim 3 , wherein the NN filter is applied to a chroma component of the video unit based on the CC-ALF being disabled, or
wherein the NN filter is applied to the chroma component of the video unit based on the CC-SAO filter being disabled.
6 . The method of claim 2 , wherein information related to the one or more filters of the NDLF tool is not included in the bitstream.
7 . The method of claim 2 , wherein information related to the one or more filters of the NDLF tool is included in the bitstream, and
wherein the information comprises an indication indicating that the one or more filters of the NDLF tool are disabled for the video unit when the NN filter is applied to the video unit.
8 . The method of claim 1 , wherein when the NN filter is applied to the video unit and information related to one or more filters of the NDLF tool is not included in the bitstream, one or more filters of the NDLF tool are determined to be disabled for the video unit.
9 . The method of claim 1 , wherein when the NN filter is applied to the video unit, all types of filters of the NDLF tool are disabled for the video unit.
10 . The method of claim 1 , wherein the video unit comprises a coding tree unit (CTU), a coding tree block (CTB), a CTU row, a CTB row, a slice, a tile, a picture, a sequence, or a subpicture.
11 . The method of claim 1 , wherein when the NN filter is disabled for the video unit, one or more filters of the NDLF tool are applied to the video unit.
12 . The method of claim 1 , wherein when one or more filters of the NDFL tool are not applied to the video unit, the NN filter is applied to the video unit.
13 . The method of claim 1 , wherein both of the NN filter and the NDFL tool are applied to the video unit, and
wherein the NN filter is applied before or after one or more filters of the NDFL tool.
14 . The method of claim 1 , wherein the NN filter is applied to the video unit after an in-loop filtering tool.
15 . The method of claim 1 , wherein the NN filter is applied to the video unit before an in-loop filtering tool,
wherein whether the in-loop filtering tool is applied to the video is determined at a slice level, a picture level, a subpicture level, or a tile level, and
wherein the in-loop filtering tool comprises a sample adaptive offset (SAO) filter, a cross-component (CC)-SAO filter, or an adaptive loop filter (ALF).
16 . The method of claim 1 , wherein the conversion comprises encoding the video into the bitstream.
17 . The method of claim 1 , wherein the conversion comprises decoding the video from the bitstream.
18 . An apparatus for processing video data comprising a processor and a non-transitory memory with instructions thereon, wherein the instructions upon execution by the processor cause the processor to:
determine, for a conversion between a video unit of a video and a bitstream of the video, interaction between a first filtering tool and a second filtering tool for being applied to the video unit, wherein the first filtering tool comprises a neural network (NN) filter, and the second filtering tool comprises a non-deep learning-based filtering (NDLF) tool; and
perform the conversion based on the determining,
wherein the interaction between the first filtering tool and the second filtering tool depends on a color format or a color component of the video unit, or
wherein the interaction between the first filtering tool and the second filtering tool depends on a profile flag, a tier flag, a level flag, or a constraint flag.
19 . A method for storing a bitstream of a video, comprising:
determining, for a video unit of the video, interaction between a first filtering tool and a second filtering tool for being applied to the video unit, wherein the first filtering tool comprises a neural network (NN) filter, and the second filtering tool comprises a non-deep learning-based filtering (NDLF) tool;
generating the bitstream based on the determining; and
storing the bitstream in a non-transitory computer-readable recording medium,
wherein the interaction between the first filtering tool and the second filtering tool depends on a color format or a color component of the video unit, or
wherein the interaction between the first filtering tool and the second filtering tool depends on a profile flag, a tier flag, a level flag, or a constraint flag.