An image decoding method performed by a decoding apparatus according to the present document comprises the steps of: obtaining image information; performing a picture management process for pictures of a decoded picture buffer (DPB) on the basis of the image information; and decoding a current picture on the basis of the pictures, wherein the image information includes an output layer set (OLS) DPB parameter index for a target OLS, and the picture management process is performed on the basis of DPB parameter information for the target OLS derived on the basis of the OLS DPB parameter index.

An image decoding method performed by a decoding apparatus according to the present document comprises the steps of: obtaining image information; performing a picture management process for pictures of a decoded picture buffer (DPB) on the basis of the image information; and decoding a current picture on the basis of the pictures, wherein the image information includes an output layer set (OLS) DPB parameter index for a target OLS, and the picture management process is performed on the basis of DPB parameter information for the target OLS derived on the basis of the OLS DPB parameter index.

Systems and methods for streaming and playing back video having a variety of resolutions, frame rates, and/or sample aspect ratios, where the video streams are encoded at one of a number of maximum bit rate levels, in accordance with embodiments of the invention are disclosed. One embodiment includes a processor, and storage containing data relating combinations of resolution and frame rates to maximum bitrates, where a plurality of resolution and frame rates that are related to the same maximum bitrate form a maximum bitrate level. In addition, an encoding application configures the processor to encode a video stream as a plurality of video streams having different resolutions and frame rates, where the target maximum bitrate used during the encoding is selected based upon the maximum bitrate levels of the resolution and frame rate combinations indicated within the data relating combinations of resolution and frame rates to maximum bitrates.

This application provides an image processing method and apparatus, a device, and a computer-readable storage medium for use in the field of computer vision. The image processing method includes the following steps: a first image device obtains a compressed image of a raw image and decompresses the compressed image to obtain a first decompressed image; the first image device determines target difference information based on difference information between the raw image and the first decompressed image; the first image device further compresses the target difference information to obtain compressed target difference information; and the first image device sends the compressed image and the compressed target difference information to a second image device. The second image device restores an image based on the received compressed image and target difference information. This application reduces a precision loss between the image restored by the second image device and the raw image.

Techniques are described for optimizing and assigning video encoding ladders.

Techniques are described for optimizing and assigning video encoding ladders.

Aspects of the disclosure provide methods and apparatuses for neural network model compression/decompression. In some examples, an apparatus for neural network model decompression includes receiving circuitry and processing circuitry. The processing circuitry decodes, from a bitstream corresponding to a representation of a neural network, at least a syntax element to be applied to multiple blocks in the neural network. Then, the processing circuitry reconstructs, from the bitstream, weight coefficients in the blocks based on the syntax element.

Aspects of the disclosure provide methods and apparatuses for neural network model compression/decompression. In some examples, an apparatus for neural network model decompression includes receiving circuitry and processing circuitry. The processing circuitry decodes, from a bitstream corresponding to a representation of a neural network, at least a syntax element to be applied to multiple blocks in the neural network. Then, the processing circuitry reconstructs, from the bitstream, weight coefficients in the blocks based on the syntax element.

A computing system may access first alpha values associated with first pixels in a first pixel region of an image and determine a bit budget for encoding the first alpha values. The computing system may then select a first alpha-encoding mode for the first alpha values to reflect a determination that the first alpha values are all fully transparent or all fully opaque, and encode the first alpha values by storing the selected first alpha-encoding mode as part of a metadata without using the bit budget to encode the first alpha values individually. The computing system may then update a record of unallocated bits available for allocation based on the bit budget unused in the encoding of the first alpha values, and allocate, based on the record of unallocated bits, bits to encode a set of alpha values different from the first alpha values.

A computing system may access first alpha values associated with first pixels in a first pixel region of an image and determine a bit budget for encoding the first alpha values. The computing system may then select a first alpha-encoding mode for the first alpha values to reflect a determination that the first alpha values are all fully transparent or all fully opaque, and encode the first alpha values by storing the selected first alpha-encoding mode as part of a metadata without using the bit budget to encode the first alpha values individually. The computing system may then update a record of unallocated bits available for allocation based on the bit budget unused in the encoding of the first alpha values, and allocate, based on the record of unallocated bits, bits to encode a set of alpha values different from the first alpha values.