High-level syntax for signaling neural networks within a media bitstream

An apparatus comprising:

• at least one processor;
• and
• at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to perform:
• encoding a media bitstream comprising one or more media units;
• encoding, within the media bitstream, at least a first enhancement information message and a third enhancement information message;
• encoding, within the first enhancement information message, a neural network or an identification of the neural network;
• and
• wherein the neural network is intended to be used to enhance or filter decoded one or more media units within a temporal scope determined with the third enhancement information message, a spatial scope determined with the third enhancement information message, or a spatio-temporal scope determined with the third enhancement information message.

A method comprising:

• receiving a media bitstream comprising one or more media units and at least a first enhancement information message, wherein the first enhancement information message comprises or identifies a neural network, and wherein the received media bitstream comprises a third enhancement information message;
• decoding the one or more media units;
• and
• using the neural network to enhance or filter the decoded one or more media units within a temporal scope determined with the third enhancement information message, a spatial scope determined with the third enhancement information message, or a spatio-temporal scope determined with the third enhancement information message.

A method comprising:

• encoding a media bitstream comprising one or more media units;
• encoding, within the media bitstream, at least a first enhancement information message and a third enhancement information message;
• encoding, within the first enhancement information message, a neural network or an identification of the neural network;
• and
• wherein the neural network is intended to be used to enhance or filter decoded one or more media units within a temporal scope determined with the third enhancement information message, a spatial scope determined with the third enhancement information message, or a spatio-temporal scope determined with the third enhancement information message.