Techniques are described for compressing and decompressing data using machine learning systems. An example process can include receiving a plurality of images for compression by a neural network compression system. The process can include determining, based on a first image from the plurality of images, a first plurality of weight values associated with a first model of the neural network compression system. The process can include generating a first bitstream comprising a compressed version of the first plurality of weight values. The process can include outputting the first bitstream for transmission to a receiver.

An attention degree estimation section in a compression coding section of a server estimates an attention degree on the basis of contents indicated by a moving image generated by an image forming section, for each of unit regions produced by dividing a frame plane. A compression coding processing section compression-codes the moving image at a compression rate varied in the frame plane according to a distribution of the attention degrees. A communication section transmits compression-coded data to an image processing apparatus.

An attention degree estimation section in a compression coding section of a server estimates an attention degree on the basis of contents indicated by a moving image generated by an image forming section, for each of unit regions produced by dividing a frame plane. A compression coding processing section compression-codes the moving image at a compression rate varied in the frame plane according to a distribution of the attention degrees. A communication section transmits compression-coded data to an image processing apparatus.

This application discloses an image processing method, including: obtaining an image; performing feature extraction on the image to obtain at least one first feature map, where the at least one first feature map includes N first feature values, and N is a positive integer; obtaining a target compression bit rate, where the target compression bit rate corresponds to M target gain values, each target gain value corresponds to one first feature value, and M is a positive integer less than or equal to N; respectively processing corresponding first feature values based on the M target gain values to obtain M second feature values; and performing quantization and entropy encoding on at least one processed first feature map to obtain encoded data, where the at least one processed first feature map includes the M second feature values. Thus, compression bit rate control can be implemented in a same compression model.

[Problem] In order to improve the likelihood of being able to continue delivery even during degradation of communication quality and improve the efficiency of use of the capability of a transmission path, even with hierarchical encoding having a small number of hierarchies, the present invention comprises: measuring the communication quality of the transmission path with a data receiving device; determining a hierarchical structure of hierarchical data on the basis of the results of measuring the communication quality; generating the hierarchical data of the hierarchical structure by hierarchical encoding of input data; and delivering delivery data of at least some hierarchies of the hierarchical data to the data receiving device.

Aspects of the disclosure provide methods and apparatuses for video encoding/decoding. In some examples, an apparatus for video decoding includes processing circuitry that determines, from a coded video bitstream, a portion of a block is subject to a zero-out operation based on a threshold number of samples being less than a number of samples in one of a height dimension or a width dimension of the block. The processing circuitry determines whether a mixture of a lossless coding mode and one or more lossy coding modes is allowed in the coded video bitstream. The processing circuitry reconstructs the block based on the one or more lossy coding modes in response to (i) the mixture of the lossless coding mode and the one or more lossy coding modes being allowed in the coded video bitstream and (ii) the portion of the block being subject to the zero-out operation.

Aspects of the disclosure provide methods and apparatuses for video encoding/decoding. In some examples, an apparatus for video decoding includes processing circuitry that determines, from a coded video bitstream, a portion of a block is subject to a zero-out operation based on a threshold number of samples being less than a number of samples in one of a height dimension or a width dimension of the block. The processing circuitry determines whether a mixture of a lossless coding mode and one or more lossy coding modes is allowed in the coded video bitstream. The processing circuitry reconstructs the block based on the one or more lossy coding modes in response to (i) the mixture of the lossless coding mode and the one or more lossy coding modes being allowed in the coded video bitstream and (ii) the portion of the block being subject to the zero-out operation.

A deblocking filter control device that controls a deblocking filter process performed on a decoded image includes: a parameter deriver configured to derive a parameter value that controls a filter strength in the deblocking filter process; and a parameter transformer configured to output a transformed parameter value by transforming the parameter value based on an input bit depth that is a bit depth of the video signal, wherein when the input bit depth is smaller than a predetermined bit depth, the parameter transformer is configured to output the transformed parameter value by adding an offset value to the parameter value and making a bit shift of a result of the addition, and the parameter transformer is configured to change the offset value based on the input bit depth.

A deblocking filter control device that controls a deblocking filter process performed on a decoded image includes: a parameter deriver configured to derive a parameter value that controls a filter strength in the deblocking filter process; and a parameter transformer configured to output a transformed parameter value by transforming the parameter value based on an input bit depth that is a bit depth of the video signal, wherein when the input bit depth is smaller than a predetermined bit depth, the parameter transformer is configured to output the transformed parameter value by adding an offset value to the parameter value and making a bit shift of a result of the addition, and the parameter transformer is configured to change the offset value based on the input bit depth.

An example apparatus includes a first frame buffer configured to store video data; a second frame buffer configured to store video data; and one or more processors configured to: reconstruct samples of a first block of a current picture of video data; store, in parallel, a compressed version of the samples of the first block of video data in the first frame buffer and an uncompressed version of the samples of the first block of video data in the second frame buffer; and responsive to determining to reconstruct a second block of the current picture of video data using intra block copy: obtain, from the second frame buffer, samples of a predictor block located in the current picture of video data, the predictor block at least partially overlapping the first block of video data; and predict, based on the obtained samples of the predictor block, samples of the second block.