Data compression and transmission method, apparatus, device, and storage medium

Abstract

The present application provides a data compression and transmission method, apparatus, device, and storage medium. A first device acquires M first data. One subdata in the first data corresponds to one first sparse matrix, and the first sparse matrix represents one corresponding subdata in the first data based on a first dictionary matrix, and the first dictionary matrix includes characteristics of the M subdata corresponding to the M first data, respectively. The first device outputs compressed data of the M first sparse matrices, where M is an integer greater than 1. In this way, effective and reliable data compression and transmission can be implemented in transmission scenarios involving large amounts of data.

Description

[Technical Field]

[0001] This application relates to the field of communications technology, and in particular to data compression and transmission methods, apparatus, devices, and storage media. [Background technology]

[0002] Currently, in some communication scenarios, such as point cloud data transmission scenarios or artificial intelligence (AI) model data (hereinafter simply referred to as AI model data) transmission scenarios, when there is a large amount of data to be transmitted between communication devices, a large number of transmission resources are occupied and transmission delays increase. In view of this, before data transmission, data compression can be performed on the data to be transmitted using a scalar quantization method or a vector quantization method, and the compressed data is then transmitted, thereby saving transmission resources and reducing transmission delays. However, compression of data to be transmitted using existing methods has a low compression rate, and the compressed data has a large data loss. Therefore, how to implement effective and reliable data compression and transmission is currently an urgent problem to be solved. Summary of the Invention

[0003] SUMMARY OF THE INVENTION Embodiments of the present application provide data compression and transmission methods, apparatus, devices, and storage media to implement efficient and reliable data compression and transmission.

[0004] According to a first aspect, an embodiment of the present application provides a data compression and transmission method, including a first device obtaining M first data, where one subdata in the first data corresponds to one first sparse matrix, the first sparse matrix represents one corresponding subdata in the first data based on a first dictionary matrix, and the first dictionary matrix includes characteristics of the M subdata respectively corresponding to the M first data, and the first device outputs compressed data of the M first sparse matrices, where M is an integer greater than 1.

[0005] According to the data compression and transmission method provided in the first aspect, dictionary learning is performed on M sub-data, each corresponding to M first data based on one first dictionary matrix, to obtain a sparse representation of each of the sub-data, thereby implementing effective and reliable data compression and transmission in transmission scenarios involving large amounts of data.

[0006] In a possible implementation, the first device outputting compressed data of the M first sparse matrices includes: the first device determining a first matrix based on the M first sparse matrices; the first device performing low-rank approximation on the first matrix to obtain first compressed data of the M first sparse matrices; and the first device outputting the first compressed data.

[0007] According to the data compression and transmission method provided in this implementation, joint compression is performed on the M first sparse matrices through low-rank approximation to further improve the compression ratio.

[0008] In a possible implementation, the first device determining the first matrix based on the M first sparse matrices includes: the first device combining the M first sparse matrices to obtain the first matrix; or the first device performing data compression on at least one of the M first sparse matrices and combining the M first sparse matrices obtained through the data compression to obtain the first matrix.

[0009] According to the data compression and transmission method provided in this implementation, when the first device combines M first sparse matrices to obtain the first matrix, there is high processing efficiency, and the first device performs data compression on at least one first sparse matrix, and then combines M first sparse matrices to obtain the first matrix, thereby further improving the compression rate.

[0010] In a possible implementation, the first device performing data compression on at least one of the M first sparse matrices includes: for one of the M first sparse matrices, the first device setting a value of a first element in the one first sparse matrix to a first value based on first position indication information, where the first position indication information indicates a position of an element in the first sparse matrix that can represent one corresponding sub-data, and the first element cannot represent one sub-data in the first data.

[0011] According to the data compression and transmission method provided in this implementation, an element capable of representing one of the corresponding sub-data is selected from the first sparse matrix, and data compression is implemented for the first sparse matrix.

[0012] According to a possible implementation, the first device performing low-rank approximation on the first matrix to obtain first compressed data of M first sparse matrices includes: the first device performing singular value decomposition on the first matrix to obtain K feature values ​​and feature vectors corresponding to the K feature values, respectively. The K feature values ​​and feature vectors corresponding to the K feature values ​​each represent the first matrix. The first device uses K0 feature values ​​among the K feature values ​​and feature vectors corresponding to the K0 feature values, respectively, as first compressed data of the M first sparse matrices.

[0013] According to the data compression and transmission method provided in this implementation, K0 characteristic values ​​and characteristic vectors respectively corresponding to the K0 characteristic values ​​are selected from the characteristic values ​​obtained through singular value decomposition, and a low-rank approximation of M first sparse matrices is implemented, thereby improving the compression rate of M first data.

[0014] In a possible implementation, the method further includes: the first device outputting second compressed data of the M first sparse matrices, the second compressed data including K1 characteristic values ​​other than the K0 characteristic values ​​among the K characteristic values, and characteristic vectors respectively corresponding to the K1 characteristic values.

[0015] According to the data compression and transmission method provided in this implementation, K1 characteristic values ​​and characteristic vectors respectively corresponding to the K1 characteristic values ​​are transmitted to complement the first compressed data, so that the second device can accurately construct M first data.

[0016] In a possible implementation, the first device outputting compressed data of the M first sparse matrices includes: the first device outputting first residual information, where the first residual information is determined based on information about the jth first sparse matrix and information about the ith first sparse matrix among the M first sparse matrices, where i is less than j, and both i and j are positive integers.

[0017] According to the data compression and transmission method provided in this implementation, joint compression is performed on the M first sparse matrices in a residual-based manner, which further improves the compression rate of the M first data.

[0018] In a possible implementation, the first device outputting the first residual information includes: the first device determining a similarity between the jth first sparse matrix and the ith first sparse matrix in the M first sparse matrices, and the first device outputs the first residual information when the similarity is less than or equal to a second similarity threshold.

[0019] According to the data compression and transmission method provided in this implementation, when the similarity between the jth first sparse matrix and the ith first sparse matrix is ​​less than or equal to the second similarity threshold, the first residual information is output, thereby avoiding the case where the jth first sparse matrix is ​​similar to the ith first sparse matrix and the first residual information is still output, and improving the compression rate of the M first data.

[0020] In a possible implementation, the information about the i-th first sparse matrix is ​​obtained by decompressing the compressed data of the i-th first sparse matrix.

[0021] According to the data compression and transmission method provided in this implementation, when the jth first sparse matrix needs to be transmitted, the first residual information is determined based on the reconstructed ith first sparse matrix and the jth first sparse matrix, so that the second device can accurately construct the jth first sparse matrix based on the first residual information.

[0022] Optionally, the first residual information includes a first residual element sequence, the first residual element sequence representing a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is less than j and both i and j are positive integers.

[0023] Optionally, the first residual information further includes second position indication information, the second position indication information indicating a position of a residual element in the residual matrix between the information for the jth first sparse matrix and the information for the ith first sparse matrix, the absolute value of which is greater than or equal to a first residual threshold.

[0024] In a possible implementation, the method further comprises: the first device outputting second residual information. The second residual information is determined based on the information about the jth first sparse matrix and the information about the ith first sparse matrix. The second residual information includes third position indication information and a second sequence of residual elements, where the third position indication information indicates residual elements in the residual matrix between the information about the jth first sparse matrix and the information about the ith first sparse matrix that are smaller than or equal to the first residual threshold and larger than or equal to the second residual threshold, and the second sequence of residual elements includes residual elements in the residual matrix that are smaller than or equal to the first residual threshold and larger than or equal to the second residual threshold.

[0025] According to the data compression and transmission method provided in this implementation, the second residual information transmitted in the increment process can complement the first residual information and enrich the residual information, so that the second device can more accurately construct the jth first sparse matrix based on the first residual information and the second residual information.

[0026] In a possible implementation, the method further comprises: k The first device receives the k-th first data Y k into a first dictionary matrix and k first sparse matrices, where k is a positive integer less than or equal to M. The first device determines M-1 first sparse matrices other than the k-th first sparse matrix based on the first dictionary matrix.

[0027] According to the data compression and transmission method provided in this implementation, when the first dictionary matrix obtained by decomposing one of the M sub-data is used for data compression and data decompression, the accuracy is higher.

[0028] In a possible implementation, for a pth first data and a qth first data among the M first data, the similarity between one sub-data in the pth first data and one sub-data in the qth first data is greater than or equal to a first similarity threshold; and both p and q are positive integers, and p is not equal to q.

[0029] According to the data compression and transmission method provided in this implementation, the sub-data in different first data are similar, so that the data compression of the M first sparse matrices is more reliable.

[0030] In a possible implementation, each of the M first data is data in M ​​time units, and one first data includes N sub-data obtained through division based on spatial positional relationship, where N is a positive integer; or the M first data is data in h time units, and the data in the h time units is sorted based on spatial positional relationship to obtain M first data, and one first data includes N sub-data, where h is a positive integer; and the similarity between the spatial position of one sub-data in the pth first data and the spatial position of one sub-data in the qth first data among the M first data is greater than or equal to a first similarity threshold.

[0031] According to the data compression and transmission method provided in this implementation, the data to be transmitted is divided to obtain M pieces of first data, so that the sparsity of the sparse matrix (e.g., M pieces of first sparse matrices) obtained through dictionary learning can be improved, and different first data are correlated, facilitating the subsequent data compression process.

[0032] In a possible implementation, the method further includes: the first device transmitting the first dictionary matrix to the second device; or the first device receiving the first dictionary matrix transmitted by the second device.

[0033] According to the data compression and transmission method provided in this implementation, the first dictionary matrix is ​​synchronized between the first device and the second device, so that the M first sparse matrices output by the first device can be used by the second device to accurately construct the M first data.

[0034] In a possible implementation, the first device outputting the first compressed data includes the first device performing a compression process on the first compressed data, the compression process including quantization and / or entropy coding, and the first device outputting the first compressed data obtained through the compression process.

[0035] According to the data compression and transmission method provided in this implementation, the compression rate of the M first sparse matrices is further improved.

[0036] In a possible implementation, the first device outputting the first residual information includes: the first device performing a compression process on the first residual information, the compression process including quantization and / or entropy coding, and the first device outputting the first residual information obtained through the compression process.

[0037] According to the data compression and transmission method provided in this implementation, the compression rate of the M first sparse matrices is further improved.

[0038] In a possible implementation, the method further includes the step of: the first device transmitting first indication information to the second device; or the first device receiving first indication information transmitted by the second device, wherein the first indication information includes: a number K0 of characteristics of the low-rank approximation; a number M of first data; a capability threshold, where the capability threshold is used to determine whether an element in the first sparse matrix can represent one corresponding sub-data in the first data; a first residual threshold, where the first residual threshold is used to determine a first residual element sequence, where the first residual element sequence represents a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is smaller than j and both i and j are positive integers; and a number M of first data corresponding to one corresponding sub-data in the first data in one of the M first sparse matrices. a proportion of elements that can represent data; data loss of the first compressed data for the M first sparse matrices, where the first compressed data is determined based on the M first sparse matrices; compression process parameters, where the compression process includes quantization and / or entropy coding, and the compression process parameters include at least one of a quantization precision, a quantization codebook, and a coding scheme; and whether to transmit first residual information, where the first residual information is determined based on information about the jth first sparse matrix and information about the i-th first sparse matrix in the M first sparse matrices.

[0039] According to the data compression and transmission method provided in this implementation, data compression and transmission are flexibly directed.

[0040] In a possible implementation, the method further includes: the first device determining, based on the M first time-frequency resources of the first data, at least one of: a capacity threshold, where the capacity threshold is used to determine whether an element in the first sparse matrix can represent one corresponding sub-data in the first data; a first residual threshold, where the first residual threshold is used to determine a first residual element sequence, where the first residual element sequence represents a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is smaller than j and both i and j are positive integers; and compression process parameters, where the compression process includes quantization and / or entropy coding, and the compression process parameters include at least one of a quantization precision, a quantization codebook, and a coding scheme.

[0041] According to the data compression and transmission method provided in this implementation, one or more thresholds and / or compression processing parameters are implicitly configured based on the first time-frequency resources of the M first data, thereby reducing the overhead of configuration signaling.

[0042] In a possible implementation, the method further comprises: the first device receiving first configuration information transmitted by the second device, the first configuration information being used to configure the first time-frequency resource.

[0043] The data compression and transmission method provided in this implementation allows for flexible configuration of transmission resources.

[0044] In a possible implementation, the method further includes the step of: the first device sending a compression and transmission request to the second device, wherein the compression and transmission request holds M data types of the first data, and the data types include point cloud data and / or artificial intelligence AI data.

[0045] In this implementation, the second device determines the time-frequency resources to be used for transmitting compressed data of the first data based on the data type, so that the first time-frequency resources configured by the second device satisfy the data type, thereby facilitating compression and transmission.

[0046] In a possible implementation, the method further includes: the first device transmitting second indication information to the second device; or the first device receiving second indication information transmitted by the second device, wherein the second indication information indicates at least one of: a number K1 of characteristics of the low-rank approximation; and a second residual threshold, where the second residual threshold is used in combination with the first residual threshold to determine a second residual element sequence, the second residual element sequence representing a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is less than j and both i and j are positive integers.

[0047] According to the data compression and transmission method provided in this implementation, data compression and transmission in the incremental transmission process are flexibly directed.

[0048] In a possible implementation, the method further includes: the first device determining a second residual threshold based on the M second time-frequency resources of the first data, the second residual threshold being used in combination with the first residual threshold to determine a second residual element sequence, the second residual element sequence representing a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is less than j and both i and j are positive integers.

[0049] The data compression and transmission method provided in this implementation reduces the configuration signaling overhead.

[0050] In a possible implementation, the method further comprises: the first device receiving second configuration information transmitted by the second device, the second configuration information being used to configure the second time-frequency resource.

[0051] According to the data compression and transmission method provided in this implementation, the transmission resources in the incremental transmission process are flexibly configured.

[0052] According to a second aspect, an embodiment of the present application provides a data compression and transmission method, comprising: a step of receiving compressed data of M first sparse matrices by a second device, the first sparse matrices representing corresponding sub-data in the first data based on a first dictionary matrix, the first dictionary matrix including characteristics of the M sub-data corresponding to the M first data respectively, and one sub-data in the first data corresponding to one first sparse matrix, and the second device outputs decompression information based on the compressed data.

[0053] In a possible implementation, the compressed data includes first compressed data, and the second device outputting decompression information based on the compressed data includes: the second device performing low-rank matrix reconstruction based on the first compressed data to obtain a first matrix; the second device determining M first sparse matrices based on the first matrix; the second device constructing M first data based on the M first sparse matrices and the first dictionary matrix; and the second device outputting the M first data.

[0054] In a possible implementation, the second device determining the M first sparse matrices based on the first matrix includes: the second device dividing the first matrix to obtain the M first sparse matrices; or the second device dividing the first matrix to obtain decompression information for the M first sparse matrices, and performing data decompression on at least one of the M first sparse matrices based on the information about the M first sparse matrices.

[0055] In a possible implementation, the second device performing data decompression on at least one of the M first sparse matrices includes: for one of the M first sparse matrices, the second device performing data decompression on the first sparse matrix based on first position indication information, where the first position indication information indicates a position of an element in the first sparse matrix that can represent one corresponding sub-data, and the first element cannot represent one sub-data in the first data.

[0056] In a possible implementation, the first compressed data includes K0 characteristic values ​​and characteristic vectors corresponding to the K0 characteristic values, respectively, and the second device performing low-rank matrix reconstruction based on the first compressed data to obtain the first matrix includes: the second device performing low-rank matrix reconstruction based on the K0 characteristic values ​​and characteristic vectors corresponding to the K0 characteristic values, respectively, to obtain the first matrix.

[0057] In a possible implementation, the compressed data further includes second compressed data, which includes K1 characteristic values ​​and characteristic vectors respectively corresponding to the K1 characteristic values, and the K1 characteristic values ​​are different from the K0 characteristic values.

[0058] In a possible implementation, the compressed data includes first residual information, the first residual information being determined based on information about the jth first sparse matrix and information about the ith first sparse matrix among the M first sparse matrices, and the second device outputting decompression information based on the compressed data includes: the second device constructing the jth first sparse matrix based on the first residual information and the ith first sparse matrix; the second device constructing M pieces of first data based on the jth first sparse matrix; and the second device outputting the M pieces of first data, where i is less than j and both i and j are positive integers.

[0059] In a possible implementation, the information about the i-th first sparse matrix is ​​obtained by decompressing the compressed data of the i-th first sparse matrix.

[0060] In a possible implementation, the first residual information includes a first residual element sequence, which represents a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is less than j and both i and j are positive integers.

[0061] In a possible implementation, the first residual information further includes second position indication information, which indicates the position of a residual element in the residual matrix between the information about the jth first sparse matrix and the information about the ith first sparse matrix, the absolute value of which is greater than or equal to a first residual threshold.

[0062] In a possible implementation, the compressed data further includes second residual information, the second residual information being determined based on the information about the jth first sparse matrix and the information about the ith first sparse matrix, the second residual information including third position indication information and a second sequence of residual elements, the third position indication information indicating residual elements in the residual matrix between the information about the jth first sparse matrix and the information about the ith first sparse matrix that are smaller than or equal to the first residual threshold and larger than or equal to the second residual threshold, and the second sequence of residual elements including residual elements in the residual matrix that are smaller than or equal to the first residual threshold and larger than or equal to the second residual threshold.

[0063] In a possible implementation, for a pth first data and a qth first data among the M first data, the similarity between one sub-data in the pth first data and one sub-data in the qth first data is greater than or equal to a first similarity threshold; and both p and q are positive integers, and p is not equal to q.

[0064] In a possible implementation, the M first data are data in M ​​time units, respectively; or the M first data are data in h time units, and the data in the h time units are sorted based on spatial positional relationships to obtain M first data, where one first data includes N sub-data, and h is a positive integer; among the M first data, the similarity between the spatial position of one sub-data in the pth first data and the spatial position of one sub-data in the qth first data is greater than or equal to a first similarity threshold.

[0065] In a possible implementation, the method further includes: the second device receiving the first dictionary matrix transmitted by the first device; or the second device transmitting the first dictionary matrix to the first device.

[0066] In a possible implementation, the method further includes: receiving, by the second device, first indication information transmitted by the first device; or transmitting, by the second device, the first indication information to the first device. The first indication information includes: a number K0 of characteristics of the low-rank approximation; a number M of first data; a capability threshold, where the capability threshold is used to determine whether an element in the first sparse matrix can represent one corresponding sub-data in the first data; a first residual threshold, where the first residual threshold is used to determine a first residual element sequence, where the first residual element sequence represents a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is smaller than j and both i and j are positive integers; and a number M of first data corresponding to one corresponding sub-data in the first data in one of the M first sparse matrices. a proportion of elements that can represent data; data loss of the first compressed data for the M first sparse matrices, where the first compressed data is determined based on the M first sparse matrices; compression process parameters, where the compression process includes quantization and / or entropy coding, and the compression process parameters include at least one of a quantization precision, a quantization codebook, and a coding scheme; and whether to transmit first residual information, where the first residual information is determined based on information about the jth first sparse matrix and information about the i-th first sparse matrix in the M first sparse matrices.

[0067] In a possible implementation, the method further comprises: the second device transmitting first configuration information to the first device, the first configuration information being used to configure the M first time-frequency resources of the first data.

[0068] In a possible implementation, the method further includes: a second device receiving a compression and transmission request sent by the first device, the compression and transmission request holding M data types of the first data, the data types including point cloud data and / or AI data.

[0069] In a possible implementation, the method further includes: receiving, by the second device, second indication information sent by the first device; or sending, by the second device, second indication information to the first device. The second indication information indicates at least one of: a number K1 of characteristics of the low-rank approximation; and a second residual threshold, where the second residual threshold is used in combination with the first residual threshold to determine a second residual element sequence, the second residual element sequence representing a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is less than j and both i and j are positive integers.

[0070] In a possible implementation, the method further includes: the first device receiving second configuration information transmitted by the second device, the second configuration information being used to configure second time-frequency resources of the M first data.

[0071] For the beneficial effects of the data compression and transmission method provided in the second aspect and possible implementations of the second aspect, please refer to the beneficial effects achieved by the first aspect and possible implementations of the first aspect, and the details will not be described again here.

[0072] According to a third aspect, an embodiment of the present application provides a communication device including: a processing module configured to acquire M first data, where one sub-data in the first data corresponds to one first sparse matrix, the first sparse matrix represents one corresponding sub-data in the first data based on a first dictionary matrix, the first dictionary matrix including characteristics of the M sub-data corresponding to the M first data respectively; and a transceiver module configured to output compressed data of the M first sparse matrices, where M is an integer greater than 1.

[0073] In a possible implementation, the processing module is further configured to determine a first matrix based on the M first sparse matrices. The first device performs low-rank approximation on the first matrix to obtain first compressed data of the M first sparse matrices. The transceiver module is configured to output the first compressed data.

[0074] In a possible implementation, the processing module is specifically configured to: combine M first sparse matrices to obtain a first matrix; or, for the first device, perform data compression on at least one of the M first sparse matrices, and combine the M first sparse matrices obtained through data compression to obtain a first matrix.

[0075] In a possible implementation, the processing module is specifically configured to: for one of the M first sparse matrices, set a value of a first element in the one first sparse matrix to a first value based on first position indication information, where the first position indication information indicates a position of an element in the first sparse matrix that can represent one corresponding sub-data, and the first element cannot represent one sub-data in the first data.

[0076] In a possible implementation, the processing module is specifically configured to perform singular value decomposition on the first matrix to obtain K feature values ​​and feature vectors respectively corresponding to the K feature values, where the K feature values ​​and the feature vectors respectively corresponding to the K feature values ​​represent the first matrix; and use K0 feature values ​​in the K feature values ​​and the feature vectors respectively corresponding to the K0 feature values ​​as first compressed data of the M first sparse matrices.

[0077] In a possible implementation, the transceiver module is further configured to output second compressed data of the M first sparse matrices, the second compressed data including K1 characteristic values ​​other than the K0 characteristic values ​​among the K characteristic values, and characteristic vectors respectively corresponding to the K1 characteristic values.

[0078] In a possible implementation, the transceiver module is specifically configured to output first residual information, which is determined based on information about a jth first sparse matrix and information about an ith first sparse matrix among the M first sparse matrices, where i is less than j, and both i and j are positive integers.

[0079] In a possible implementation, the processing module is further configured to determine a similarity between the jth first sparse matrix and the ith first sparse matrix among the M first sparse matrices; and the transceiver module outputs the first residual information when the similarity is less than or equal to a second similarity threshold.

[0080] In a possible implementation, the information about the i-th first sparse matrix is ​​obtained by decompressing the compressed data of the i-th first sparse matrix.

[0081] In a possible implementation, the first residual information includes a first residual element sequence, which represents a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is less than j and both i and j are positive integers.

[0082] In a possible implementation, the first residual information further includes second position indication information, which indicates an absolute value that is greater than or equal to the absolute value in the residual matrix between the information about the jth first sparse matrix and the information about the ith first sparse matrix.

[0083] In a possible implementation, the transceiver module is further configured to output second residual information, the second residual information being determined based on the information about the jth first sparse matrix and the information about the ith first sparse matrix, the second residual information including third position indication information and a second sequence of residual elements, the third position indication information indicating residual elements in the residual matrix between the information about the jth first sparse matrix and the information about the ith first sparse matrix that are smaller than or equal to the first residual threshold and larger than or equal to the second residual threshold, and the second sequence of residual elements including residual elements in the residual matrix that are smaller than or equal to the first residual threshold and larger than or equal to the second residual threshold.

[0084] In a possible implementation, the processing module further comprises: k The first device receives the k-th first data Y k into a first dictionary matrix and k first sparse matrices, where k is a positive integer less than or equal to M. The first device determines M-1 first sparse matrices other than the k-th first sparse matrix based on the first dictionary matrix.

[0085] In a possible implementation, for a pth first data and a qth first data among the M first data, the similarity between one sub-data in the pth first data and one sub-data in the qth first data is greater than or equal to a first similarity threshold; and both p and q are positive integers, and p is not equal to q.

[0086] In a possible implementation, the M first data are data in M ​​time units, respectively; or the M first data are data in h time units, and the data in the h time units are sorted based on spatial positional relationships to obtain M first data, where one first data includes N sub-data, and h is a positive integer; among the M first data, the similarity between the spatial position of one sub-data in the pth first data and the spatial position of one sub-data in the qth first data is greater than or equal to a first similarity threshold.

[0087] In a possible implementation, the transceiver module is further configured to: transmit the first dictionary matrix to the second device; or receive the first dictionary matrix transmitted by the second device.

[0088] In a possible implementation, the transceiver module is specifically configured to: perform a compression process on the first compressed data, where the compression process includes quantization and / or entropy coding; and output the first compressed data obtained through the compression process.

[0089] In a possible implementation, the transceiver module is specifically configured to: perform a compression process on the first residual information, where the compression process includes quantization and / or entropy coding; and output the first residual information obtained through the compression process.

[0090] In a possible implementation, the following is further included: the transceiver module is further configured to: transmit first indication information to the second device; or receive first indication information transmitted by the second device, the first indication information including: a number K0 of characteristics of the low-rank approximation; a number M of first data; a capability threshold, where the capability threshold is used to determine whether an element in the first sparse matrix can represent one corresponding sub-data in the first data; a first residual threshold, where the first residual threshold is used to determine a first residual element sequence, where the first residual element sequence represents a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is smaller than j and both i and j are positive integers; and a number M of first data corresponding to one corresponding sub-data in the first data in one of the M first sparse matrices. a proportion of elements that can represent data; data loss of the first compressed data for the M first sparse matrices, where the first compressed data is determined based on the M first sparse matrices; compression process parameters, where the compression process includes quantization and / or entropy coding, and the compression process parameters include at least one of a quantization precision, a quantization codebook, and a coding scheme; and whether to transmit first residual information, where the first residual information is determined based on information about the jth first sparse matrix and information about the i-th first sparse matrix in the M first sparse matrices.

[0091] In a possible implementation, the processing module is further configured to determine, based on the M first time-frequency resources of the first data: a capacity threshold, where the capacity threshold is used to determine whether an element in the first sparse matrix can represent one corresponding sub-data in the first data; a first residual threshold, where the first residual threshold is used to determine a first residual element sequence, where the first residual element sequence represents a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is smaller than j and both i and j are positive integers; and compression processing parameters, where the compression processing includes quantization and / or entropy coding, and the compression processing parameters include at least one of a quantization precision, a quantization codebook, and a coding scheme.

[0092] In a possible implementation, the transceiver module is further configured to receive first configuration information transmitted by the second device, the first configuration information being used to configure the first time-frequency resource.

[0093] In one possible implementation, the transceiver module is further configured to send a compression and transmission request to the second device, wherein the compression and transmission request holds M data types of the first data, and the data types include point cloud data and / or AI data.

[0094] In a possible implementation, the transceiver module is further configured to: transmit second instruction information to the second device; or receive second instruction information transmitted by the second device, wherein the second instruction information indicates at least one of: a number K1 of characteristics of the low-rank approximation; and a second residual threshold, where the second residual threshold is used in combination with the first residual threshold to determine a second residual element sequence, the second residual element sequence representing a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is less than j and both i and j are positive integers.

[0095] In a possible implementation, the processing module is further configured to determine a second residual threshold based on the M second time-frequency resources of the first data, the second residual threshold being used in combination with the first residual threshold to determine a second residual element sequence, the second residual element sequence representing a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is less than j and both i and j are positive integers.

[0096] In a possible implementation, the transceiver module is further configured to receive second configuration information transmitted by the second device, the second configuration information being used to configure the second time-frequency resource.

[0097] For beneficial effects of the communication device provided in the third aspect and possible implementations of the third aspect, please refer to the beneficial effects achieved by the first aspect and possible implementations of the first aspect, and details will not be described again here.

[0098] According to a fourth aspect, an embodiment of the present application provides a communication device including: a transceiver module configured to receive compressed data of M first sparse matrices, where the first sparse matrices represent corresponding sub-data in the first data based on a first dictionary matrix, the first dictionary matrix including characteristics of the M sub-data corresponding to the M first data respectively, and one sub-data in the first data corresponds to one first sparse matrix; and a processing module configured to output decompression information based on the compressed data.

[0099] In a possible implementation, the compressed data includes first compressed data, and the processing module is specifically configured to: perform low-rank matrix reconstruction based on the first compressed data to obtain a first matrix; determine M first sparse matrices based on the first matrix; construct M first data based on the M first sparse matrices and the first dictionary matrix; and output the M first data to the second device.

[0100] In a possible implementation, the processing module is specifically configured to: divide the first matrix to obtain M first sparse matrices; or divide the first matrix to obtain decompression information for the M first sparse matrices, and perform data decompression on at least one of the M first sparse matrices based on the information about the M first sparse matrices.

[0101] In a possible implementation, the processing module is specifically configured to, for one of the M first sparse matrices, perform data decompression on the first sparse matrix based on first position indication information, where the first position indication information indicates a position of an element in the first sparse matrix that can represent one corresponding sub-data, and the first element cannot represent one sub-data in the first data.

[0102] In a possible implementation, the first compressed data includes K0 characteristic values ​​and characteristic vectors corresponding to the K0 characteristic values, respectively, and the second device performing low-rank matrix reconstruction based on the first compressed data to obtain the first matrix includes: the second device performing low-rank matrix reconstruction based on the K0 characteristic values ​​and characteristic vectors corresponding to the K0 characteristic values, respectively, to obtain the first matrix.

[0103] In a possible implementation, the compressed data further includes second compressed data, which includes K1 characteristic values ​​and characteristic vectors respectively corresponding to the K1 characteristic values, and the K1 characteristic values ​​are different from the K0 characteristic values.

[0104] In a possible implementation, the compressed data includes first residual information, the first residual information being determined based on information about the jth first sparse matrix and information about the ith first sparse matrix among the M first sparse matrices, and the second device outputting decompression information based on the compressed data includes: the second device constructing the jth first sparse matrix based on the first residual information and the ith first sparse matrix; the second device constructing M pieces of first data based on the jth first sparse matrix; and the second device outputting the M pieces of first data, where i is less than j and both i and j are positive integers.

[0105] In a possible implementation, the information about the i-th first sparse matrix is ​​obtained by decompressing the compressed data of the i-th first sparse matrix.

[0106] In a possible implementation, the first residual information includes a first residual element sequence, which represents a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is less than j and both i and j are positive integers.

[0107] In a possible implementation, the first residual information further includes second position indication information, which indicates the position of a residual element in the residual matrix between the information about the jth first sparse matrix and the information about the ith first sparse matrix, the absolute value of which is greater than or equal to a first residual threshold.

[0108] In a possible implementation, the compressed data further includes second residual information, the second residual information being determined based on the information about the jth first sparse matrix and the information about the ith first sparse matrix, the second residual information including third position indication information and a second sequence of residual elements, the third position indication information indicating residual elements in the residual matrix between the information about the jth first sparse matrix and the information about the ith first sparse matrix that are smaller than or equal to the first residual threshold and larger than or equal to the second residual threshold, and the second sequence of residual elements including residual elements in the residual matrix that are smaller than or equal to the first residual threshold and larger than or equal to the second residual threshold.

[0109] In a possible implementation, for a pth first data and a qth first data among the M first data, the similarity between one sub-data in the pth first data and one sub-data in the qth first data is greater than or equal to a first similarity threshold; and both p and q are positive integers, and p is not equal to q.

[0110] In a possible implementation, the M first data are data in M ​​time units, respectively; or the M first data are data in h time units, and the data in the h time units are sorted based on spatial positional relationships to obtain M first data, where one first data includes N sub-data, and h is a positive integer; among the M first data, the similarity between the spatial position of one sub-data in the pth first data and the spatial position of one sub-data in the qth first data is greater than or equal to a first similarity threshold.

[0111] In a possible implementation, the transceiver module is further configured to: receive the first dictionary matrix transmitted by the first device; or transmit the first dictionary matrix to the first device.

[0112] In a possible implementation, the transceiver module is further configured to: receive first indication information transmitted by the first device; or transmit first indication information to the first device, wherein the first indication information includes: a number K0 of characteristics of the low-rank approximation; a number M of first data; a capability threshold, where the capability threshold is used to determine whether an element in the first sparse matrix can represent one corresponding sub-data in the first data; a first residual threshold, where the first residual threshold is used to determine a first residual element sequence, where the first residual element sequence represents a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is smaller than j and both i and j are positive integers; and a number M of first data corresponding to one corresponding sub-data in the first data in one of the M first sparse matrices. a proportion of elements that can represent data; data loss of the first compressed data for the M first sparse matrices, where the first compressed data is determined based on the M first sparse matrices; compression process parameters, where the compression process includes quantization and / or entropy coding, and the compression process parameters include at least one of a quantization precision, a quantization codebook, and a coding scheme; and whether to transmit first residual information, where the first residual information is determined based on information about the jth first sparse matrix and information about the i-th first sparse matrix in the M first sparse matrices.

[0113] In a possible implementation, the transceiver module is further configured to send first configuration information to the first device, the first configuration information being used to configure the first time-frequency resources of the M first data.

[0114] In a possible implementation, the transceiver module is further configured to receive a compression and transmission request sent by the first device, wherein the compression and transmission request holds M first data data types, and the data types include point cloud data and / or artificial intelligence (AI) data.

[0115] In a possible implementation, the transceiver module is further configured to: receive second indication information sent by the first device; or send second indication information to the first device, where the second indication information indicates at least one of: a number K1 of characteristics of the low-rank approximation; and a second residual threshold, where the second residual threshold is used in combination with the first residual threshold to determine a second residual element sequence, and the second residual element sequence represents a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is less than j and both i and j are positive integers.

[0116] In a possible implementation, the transceiver module is further configured to receive second configuration information sent by the second device, the second configuration information being used to configure the second time-frequency resources of the M first data.

[0117] For beneficial effects of the communication device provided in the fourth aspect and possible implementations of the fourth aspect, please refer to the beneficial effects achieved by the first aspect and possible implementations of the first aspect, and details will not be described again here.

[0118] According to a fifth aspect, an embodiment of the present application provides a communications device comprising a processor configured to perform the method of the first aspect, the second aspect or possible implementations by executing a computer program or by using logic circuitry.

[0119] In a possible implementation, a memory is further included, the memory being configured to store a computer program.

[0120] In a possible implementation, a communication interface is further included, the communication interface being configured to input and / or output signals.

[0121] According to a sixth aspect, an embodiment of the present application provides a communications device comprising a processor and a memory, the memory configured to store a computer program, and the processor configured to invoke and execute the computer program stored in the memory to perform the method of the first aspect, the second aspect, or possible implementations.

[0122] According to a seventh aspect, embodiments of the present application provide a chip including a processor configured to retrieve computer instructions from a memory and execute the computer instructions to cause a device in which the chip is installed to perform a method of the first aspect, the second aspect, or possible implementations thereof.

[0123] According to an eighth aspect, an embodiment of the present application provides a computer-readable storage medium configured to store computer program instructions, the computer program causing a computer to perform the method of the first aspect, the second aspect, or possible implementations.

[0124] According to a ninth aspect, embodiments of the present application provide a computer program product comprising computer program instructions, the computer program instructions causing a computer to perform the method of the first aspect, the second aspect or possible implementations.

[0125] According to a tenth aspect, an embodiment of the present application provides a computer program, the computer program causing a computer to perform the method of the first aspect, the second aspect or possible implementations. [Brief explanation of the drawings]

[0126] [Figure 1] 1 is a diagram of the architecture of a mobile communication system to which embodiments of the present application apply;

[0127] [Figure 2] FIG. 1 is a diagram of a dictionary learning framework according to an embodiment of the present application.

[0128] [Figure 3] 1 is a schematic interaction flowchart of a data compression and transmission method according to an embodiment of the present application;

[0129] [Figure 4a] FIG. 1 is a diagram of data partitioning according to an embodiment of the present application.

[0130] [Figure 4b] FIG. 10 is a diagram of another data division according to an embodiment of the present application.

[0131] [Figure 5] FIG. 10 is a diagram of another data division according to an embodiment of the present application.

[0132] [Figure 6] FIG. 1 is an illustration of data compression of a sparse matrix according to an embodiment of the present application.

[0133] [Figure 7] 4 is a schematic interaction flowchart of another data compression and transmission method according to an embodiment of the present application;

[0134] [Figure 8] 1 is a schematic block diagram of a communication device according to an embodiment of the present application;

[0135] [Figure 9] FIG. 2 is another schematic block diagram of a communication device according to an embodiment of the present application; DETAILED DESCRIPTION OF THE INVENTION

[0136] The technical solutions of the present application are described below with reference to the accompanying drawings.

[0137] FIG. 1 is a diagram illustrating the architecture of a mobile communication system to which an embodiment of the present application is applied. As shown in FIG. 1, the mobile communication system includes a core network device 110, a network device 120, and at least one terminal device (terminal device 130 and terminal device 140 shown in FIG. 1). The terminal device is connected to the network device wirelessly, and the network device is connected to the core network device wirelessly or wired. The core network device and the network device may be different physical devices independent of each other; the functions of the core network device and the logical functions of the network device may be integrated into the same physical device; or some functions of the core network device and some functions of the network device may be integrated into one physical device. The terminal device may be located at a fixed location or may be mobile. FIG. 1 is merely a diagram. The communication system may further include other network devices, for example, wireless relay devices and wireless backhaul devices, which are not shown in FIG. 1. The number of core network devices, network devices, and terminal devices included in the mobile communication system is not limited in the embodiments of the present application.

[0138] In the embodiments of the present application, the network device may be any device with a wireless transceiver function, including, but not limited to, an evolved NodeB (eNB), a home NodeB (e.g., home evolved NodeB or home NodeB, HNB), a baseband unit (BBU), an access point (AP) in a wireless fidelity (Wi-Fi®) system, a wireless relay node, a wireless backhaul node, a transmission point (TP) or a transmission and reception point (TRP), a mobile switching center, a device performing base station functions in device-to-device (D2D), vehicle-to-everything (V2X), machine-to-machine (M2M) communications, and unmanned aerial vehicle communications, and a non-terrestrial network, The antenna element may be a network device (which may be deployed on a high-altitude platform, satellite, or high-altitude aircraft) in a 5G (New Telecommunications Network) communication system, a gNB in ​​a 5G system, or one or a group of antenna panels (including multiple antenna panels) of a base station in a 5G system, or a network node forming a gNB or a transmission point, such as a BBU or a distributed unit (DU). This is not specifically limited in the embodiments of the present application.

[0139] In some deployments, a gNB may include a centralized unit (CU) and a DU. The CU and DU may each implement some functions of the gNB, and the CU and DU may communicate with each other over an F1 interface. The gNB may also include an active antenna unit (AAU). The AAU may implement some physical layer processing functions, radio frequency processing, and active antenna-related functions.

[0140] It may be understood that the network device may be a device including one or more of a CU node, a DU node, and an AAU node. In addition, the CU may be classified as a network device in an access network (radio access network, RAN), or the CU may be classified as a network device in a core network (CN). This is not a limitation in the present application.

[0141] In embodiments of the present application, a terminal device may also be referred to as user equipment (UE), access terminal, subscriber unit, subscriber station, mobile station, remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent, or user equipment.

[0142] The terminal device may be a device that provides voice / data connectivity to a user, such as a handheld device or an in-vehicle device with wireless connectivity. Currently, some examples of terminals are mobile phones, tablet computers (pads), computers with wireless transceiver capabilities (e.g., notebook computers or palmtop computers), unmanned aerial vehicles, customer-premises equipment (CPE), point of sale (POS) machines, mobile internet devices (MIDs), virtual reality (VR) devices, augmented reality (AR) devices, wireless terminals in industrial control, wireless terminals in self driving, wireless terminals in remote medical, wireless terminals in smart grids, wireless terminals in transportation safety, wireless terminals in smart cities, wireless terminals in smart homes, cellular telephones, cordless telephones, session initiation protocol (SIP) telephones, wireless local loop (WLL) stations, personal digital assistants (PDAs), and so on. The device may be a PDA (registered trademark), a handheld device with wireless communication capabilities, a computing device, another processing device connected to a wireless modem, an in-vehicle device, a wearable device, a terminal device in a 5G network, a terminal device in a system evolved after 5G, or the like.

[0143] The network device and the terminal device may communicate with each other through a licensed spectrum, or through an unlicensed spectrum, or through both the licensed spectrum and the unlicensed spectrum. The network device and the terminal device may communicate with each other over a spectrum below 6G, over a spectrum above 6G, or over both the spectrum below 6G and the spectrum above 6G. The spectrum resources used between the network device and the terminal device are not limited in the embodiments of the present application.

[0144] The specific configuration of the network device and the terminal device is not limited in this application.

[0145] The communication method provided in the present application may be applied to various communication systems, such as a Long Term Evolution (LTE) system, a 5G mobile communication system, and a 6G mobile communication system that is an evolution of 5G. The 5G mobile communication system or the 6G mobile communication system may include non-standalone (NSA) networking and / or standalone (SA) networking.

[0146] The communication methods provided herein may also be applied to machine type communication (MTC), Long Term Evolution-machine (LTE-M), device-to-device (D2D) networks, machine-to-machine (M2M) networks, internet of things (IoT) networks, or other networks.

[0147] Point cloud data is a set of spatial points in a three-dimensional coordinate system. For example, data is collected by using a machine vision sensor and recorded in the form of points, and each spatial point includes three-dimensional coordinates and may also include color information (RGB), location information, reflection intensity information, and the like. In a point cloud scenario, a terminal device (e.g., 130 and / or 140 in FIG. 1) may collect data by using a sensor and transmit the collected data to a network device (e.g., 120 in FIG. 1), which may reconstruct the point cloud data.

[0148] Federated learning is a distributed machine learning technique that performs distributed model training among multiple data sources with local data, and builds a global model based on virtual aggregate data by exchanging only model parameters or intermediate results without exchanging local individual or sample data, thereby balancing data privacy protection, data sharing, and computing. The model parameters, AI gradients, or intermediate results transmitted in the federated learning process may be referred to as AI model data. Of course, AI model data is merely a possible name, and the name is not limited in this application. For example, AI model data may also be referred to as model data or AI data. In a federated learning scenario, a terminal device (e.g., 130 and / or 140 in FIG. 1 ) may transmit locally updated AI model data to a network device (e.g., 120 in FIG. 1 ) to update the global AI model data.

[0149] The data volumes of point cloud data and AI model data are large. Typically, point cloud data may contain hundreds of thousands of spatial points, and AI model data may reach tens of millions of dimensions.

[0150] Currently, there is no effective and reliable compression and transmission solution for communication scenarios in which a large amount of data is transmitted, such as point cloud scenarios and federated learning scenarios. In this case, in an embodiment of the present application, for multiple pieces of data to be transmitted (for example, M first data, where M is an integer greater than 1), dictionary learning is performed on one sub-data in each piece of data based on the same dictionary matrix to obtain a sparse representation of each sub-data, thereby implementing effective and reliable data compression of data to be transmitted involving a large amount of data.

[0151] In the above example, the uplink transmission of point cloud data and AI model data is used only as an example for explanation and should not be considered as any limitation to the present application. For example, in a point cloud scenario, terminal device A (e.g., 130 in FIG. 1) may transmit point cloud data to terminal device B (e.g., 140 in FIG. 1) through forwarding by a network device (e.g., 120 in FIG. 1); or terminal device A may transmit point cloud data to terminal device B through a sidelink. In another example, in a federated learning scenario, a network device (e.g., 120 in FIG. 1) may transmit locally updated AI model data to terminal devices (e.g., 130 and / or 140 in FIG. 1).

[0152] The data compression and transmission method provided in the embodiments of the present application will be mainly described by using an example in which the method is applied to a communication system. However, this is not limited to this application. For example, the data compression and transmission method provided in the present application can also be applied to any electronic device with processing capabilities. The electronic device can be any terminal device or server. The electronic device can implement data compression based on the method provided in the embodiments of the present application, output compressed data, and transmit the compressed data to another electronic device, for example.

[0153] To facilitate understanding of the present application, dictionary learning is first described as an example.

[0154] 2 is a diagram of a dictionary learning framework according to an embodiment of the present application. As shown in FIG. 2, source data Y may be represented by a matrix with L rows and Q columns, also referred to as a source matrix, and the source data Y is decomposed into a dictionary matrix D and a sparse code matrix X through a numerical iterative process of dictionary learning, i.e., Y=DX. Each element in the matrix representation of the source data Y may be a floating-point number.

[0155] The dictionary matrix D, also referred to as a base, may be a matrix with L rows by L columns, and each column in the dictionary matrix D may be referred to as a base vector. The dictionary matrix D includes characteristics of the source data. For example, the characteristics of the source data are represented through each base vector.

[0156] The sparse matrix X may represent source data based on the dictionary matrix D. The sparse matrix X represents characteristics of the source data by using a combination of basis vectors in the dictionary matrix D, for example, a combination obtained by weighting each basis vector in the dictionary matrix D. The sparse matrix X is a matrix with L rows and Q columns. Each column of a vector in the sparse matrix X may represent each column of a vector in the source matrix Y. For example, the first column l1 of a vector in the sparse matrix X may represent y1 in the source matrix Y.

[0157] Each row of the vector in the sparse matrix X corresponds to each basis vector in the dictionary matrix D, and the value of each row in the column of the sparse matrix X represents the weight of each basis vector in the dictionary matrix D. For example, the first column x1 of the vector in the sparse matrix X represents y1 in the source data Y, and the elements in the row of the first column of the vector in the sparse matrix X are sequentially 11 , x 12 , ..., x 1L In this case, x 11 is the weight of the first basis vector d1 in the dictionary matrix used when the sparse matrix X represents the source matrix Y.

[0158] The elements in the sparse matrix X may be represented or stored in a coordinate format, and each element may include row information, column information, and the element value of the element in the sparse matrix X. Therefore, the elements in the sparse matrix X may also be referred to as coordinates.

[0159] A larger number of elements whose value is 0 in the sparse matrix X indicates that a smaller resource is occupied by information that has a low correlation with the learning task, and therefore the sparse matrix X has a good ability to represent the source data Y while reducing the storage resource overhead and the transmission resource overhead. In this case, the sparse performance of the sparse matrix X is considered to be good.

[0160] The following describes a communication method provided in an embodiment of the present application with reference to the accompanying drawings.

[0161] For ease of understanding and description, it should be understood that the following describes the method provided in the embodiments of the present application by using a first device as an execution subject. The first device may be any terminal device in the communication system shown in FIG. 1, such as the terminal device 130 or the terminal device 140, or the first device may be the network device 120 in the communication system shown in FIG. 1. In some embodiments, the interaction between the first device and the second device is used as an example to describe the method provided in the embodiments of the present application. When the data compression and transmission method provided in the embodiments of the present application is applied to uplink transmission, the first device may be any terminal device in the communication system shown in FIG. 1, such as the terminal device 130 or the terminal device 140, and the second device may be the network device 120 in the communication system shown in FIG. 1. When the data compression and transmission method provided in the embodiments of the present application is applied to downlink transmission, the first device may be the network device 120 in the communication system shown in FIG. 1, and the second device may be any terminal device in the communication system shown in FIG. 1, such as the terminal device 130 or the terminal device 140. When the data compression and transmission method provided in the embodiments of the present application is applied to sidelink transmission, the first communication device may be any terminal device in the communication system shown in Figure 1, for example, terminal device 130, and the second communication device may be any terminal device in the communication system shown in Figure 1 other than the first communication device, for example, terminal device 140.

[0162] It should be further understood that this should not constitute any limitation on the entity that performs the method provided in the present application. Any device can provide a service to the entity that performs the method provided in the embodiments of the present application, provided that the device can perform the method provided in the embodiments of the present application by executing a program including the code of the method provided in the embodiments of the present application. For example, any one of the above communication devices can be implemented as a terminal device, or as a component in a terminal device, such as a chip, a chip system, or another functional module that can call a program to execute the program. Any one of the above communication devices can be implemented as a network device, or as a component in a network device, such as a chip, a chip system, or another functional module that can call a program to execute the program.

[0163] 3 is a schematic interaction flowchart of a data compression and transmission method according to an embodiment of the present application. Referring to FIG. 3, the method 200 includes some or all of the following processes:

[0164] S210: A first device acquires M first data, where one sub-data in the first data corresponds to one first sparse matrix, the first sparse matrix represents one corresponding sub-data in the first data based on a first dictionary matrix, and the first dictionary matrix includes characteristics of the M sub-data corresponding to the M first data respectively.

[0165] S220: The first device transmits compressed data of the M first sparse matrices to the second device, and correspondingly, the second device receives the compressed data of the M first sparse matrices transmitted by the first device.

[0166] S230: The second device outputs decompressed data based on the compressed data.

[0167] The M pieces of first data may be M pieces of sub-data in the data to be transmitted. In other words, the M pieces of first data may be obtained by dividing the data to be transmitted. Therefore, the M pieces of first data may be similar in time dimension and / or space dimension. Of course, this is not limited in the present application. For example, the M pieces of first data may be any M pieces of data in the data to be transmitted.

[0168] When the first data to be transmitted is used as the source data Y and data compression is performed based on dictionary learning, it should be noted that when the amount of the first data is large, the source data Y is widely distributed in space, making it difficult to determine the dictionary matrix and obtain a sparse matrix with good sparsity. As a result, the data processing process of obtaining the dictionary matrix D and the sparse matrix X through decomposition is complex and has a long delay. In view of this, in this embodiment of the present application, the first device may divide the data to be transmitted into M first data and perform dictionary learning for each of the first data. Furthermore, if the amount of the first data is still large, the first data may be further divided to help further improve the sparsity of the sparse matrix. In this embodiment of the present application, there may be, but are not limited to, the following two data division methods. The following describes the two data division methods by using examples with reference to Figures 4a and 4b.

[0169] Method 1: Data decomposition is performed in the time dimension.

[0170] Referring to FIG. 4a, in Scheme 1, the data to be transmitted may include data in M ​​time units, where each data in each time unit is used as one first data, and each time unit may include N sub-data, where N is a positive integer. The time unit may be one or more data frames, one or more slots, or the like, and should not be understood as the smallest time unit. Adjacent time units in the M time units may be two consecutive time units or two time units with a time interval. The time interval between the two time units must be less than or equal to a predetermined time interval, thereby ensuring the similarity between the M time units.

[0171] It should be noted that in Scheme 1, each of the M first data includes one subdata related to a spatial dimension. Specifically, the similarity between any two spatial locations of the M subdata corresponding to the M first data is greater than or equal to a first similarity threshold. For example, the similarity between the spatial location of one subdata in the pth first data and the spatial location of one subdata in the qth first data is greater than or equal to a first similarity threshold, where both p and q are positive integers, and p is not equal to q. As shown in FIG. 4a, the first subdata Y in the first first data Y1 is (1,1) and the first sub-data Y in the second first data Y2 (2,1) the similarity between the first sub-data Y in the second first data Y2 is greater than or equal to the first similarity threshold; (2,1) and the first sub-data Y in the third first data Y3 (3,1) The similarity between the (M-1)th first data Y is greater than or equal to the first similarity threshold; M-1 The first subdata Y in (M-1,1) and M's first data Y M The first subdata Y in (M,1) Similarly, in FIG. 4a, the similarity between the second sub-data Y in the first data Y1 is greater than or equal to the first similarity threshold. (1,2)and the second sub-data Y in the second first data Y2 (2,2) the similarity between the second sub-data Y in the second first data Y2 is greater than or equal to the first similarity threshold; (2,2) and the second sub-data Y in the third first data Y3 (3,2) The similarity between the (M-1)th first data Y is greater than or equal to the first similarity threshold; M-1 The second subdata Y in (M-1,2) and M's first data Y M The second subdata Y in (M,2) The similarity between the first data and the subdata is greater than or equal to a first similarity threshold. Other subdata in the first data may also have the above similarity. For brevity, details will not be described again. However, this is not a limitation in the present application. For example, any first data may include at least one subdata that does not have the above similarity with the subdata in other first data. In addition, the number of subdata included in the M first data may vary. For ease of explanation, the following provides an explanation by using an example in which all n-th subdata in the M first data are correlated. Here, n is a positive integer less than or equal to N.

[0172] In Scheme 1, for any of the M first data (for example, the mth first data, where m is a positive integer less than or equal to M), the first data may be a matrix or may be expressed in the form of a matrix. The first device may store the first data in a matrix Y with L rows and Q columns. m For example, L is equal to the dimension of the spatial points, i.e., L=3, and Q is equal to the total number of spatial points in the point cloud data. In another example, the first data is AI model data, and L may be a preset value, for example, a value obtained through experiments that enables better sparse performance of the sparse matrix. In this case,

number

number

[0173] The first device may divide the first data into N sub-data, for example, based on a spatial relationship between all data in the first data, and each of the N sub-data may be a matrix.

[0174] Referring to FIG. 5, the mth first data Y m Q column (e.g., y1~y Q ) can be divided into N sub-data. For example, y1 to y K is the first sub-data Y (m,1) It is classified as y K+1 ~y 2K is the second subdata Y (m,2) , ... and y Q-K+1 ~y Q is the Nth subdata Y (m,N=Q / K) 5 is described by using only an example in which the number of columns of the matrices formed by all of the N subdata is the same. However, this is not limited to this example. For example, the number of columns of the matrices of at least two of the N subdata is different. When the number of columns of the matrices formed by all of the subdata is the same, the complexity of data processing of the first communication device and the second communication device can be reduced.

[0175] For example, the first data is point cloud data. The first device may divide the first data into N subdata based on the distances between all spatial points in the first data in Euclidean space. For example, P spatial points may be determined in ascending order of the distances between the spatial points, where P is a preset value, and the P spatial points are classified into one subdata. Alternatively, spatial points within a preset range are classified into subdata. For example, spatial points on a circle whose center is the spatial point and whose radius is R are classified into one subdata, where R is a preset value. In different first data, subdata related in spatial dimensions may be data of the same or similar spatial points in Euclidean space. For example, the nth subdata Y in the first first data Y is classified into one subdata. (1,n) , the n-th sub-data Y in the second first data Y2 (2,n) , ..., and the Nth sub-data Y in the Mth first data (M,N) are all data from the same spatial point (or several identical spatial points) at different time units.

[0176] For example, the first data is AI model data. The first device may divide the first data into M subdata based on the neural network layer in which all data in the first data is located. For example, data in the same neural network layer or multiple adjacent neural network layers are classified into one subdata. If the number of data in the same neural network layer exceeds the data amount limit of one subdata, the data in the same neural network layer may be classified into two or more adjacent subdata. Neural network layers include, but are not limited to, fully connected layers, convolutional layers, pooling layers, and the like. In different first data, subdata related in the spatial dimension may be data in the same or similar neural network layer. For example, the nth subdata Y in the first first data Y is (1,n) , the n-th sub-data Y in the second first data Y2 (2,n) , ..., and the Nth sub-data Y in the Mth first data (M,N)are all data of the same neural network layer (or similar neural network layers) at different time units.

[0177] It should be further understood that whether the subdata are similar can be determined based on the similarity of their spatial locations or based on the similarity in the time dimension. For example, if the difference between the collection time of one subdata in the pth first data and the collection time of one subdata in the qth first data is less than a predetermined time difference, the similarity between the two subdata is greater than or equal to a first similarity threshold. The similarity between the subdata can also be expressed as a difference less than or equal to a predetermined difference threshold, and the difference between the subdata can be determined based on the residual difference between the subdata.

[0178] Method 2: Data partitioning is performed in the spatial dimension.

[0179] 4b, in Scheme 2, the data to be transmitted may be data in h time units, where h is a positive integer, and the M first data may be obtained through division by sorting the data to be transmitted in the h time units based on spatial location relationships. The time units are described in Scheme 1. For brevity, the details will not be described again.

[0180] The data to be transmitted may be a matrix or may be expressed in the form of a matrix. The first device may arrange the data to be transmitted into a matrix Y with L rows and Q columns. Furthermore, the first device may divide the data to be transmitted in h time units into N' sub-data. For example, the first data is divided into N' sub-data based on the relationship between all data in the first data in space. Each of the sub-data may be a matrix. For details, please refer to the description of the first data in Scheme 1. For brevity, the details will not be described again.

[0181] The first device may classify N' sub-data into M first data by arranging the sub-data one by one in a round-robin manner. For example, y1 in the data to be transmitted is arranged in the first first data Y1, and the first sub-data Y in the first first data Y1 is arranged in the first first data Y2. (1,1) y2 is placed in the second first data Y2, and the first sub-data Y (2,1) used as,...,y M is the Mth first data Y M and the first data Y M The first subdata Y in (M,1) is used as y M+1 is placed in the first data Y1, and the second sub-data Y (1,2) and so on until all N' sub-data in the data to be transmitted are placed into M first data.

[0182] Therefore, in FIG. 4b, the first sub-data Y in the first data Y1 (1,1) , the first sub-data Y in the second first data Y2 (2,1) , ..., and M first data Y M The first subdata Y in (M,1) are correlated in the spatial dimension. Specifically, the first sub-data Y in the first data Y1 (1,1) , and the first sub-data Y in the second first data Y2 (2,1) the similarity between the first sub-data Y in the second first data Y2 is greater than or equal to the first similarity threshold; (2,1) , and the first sub-data Y in the third first data Y3 (3,1) The similarity between the (M-1)th first data Y is greater than or equal to the first similarity threshold; M-1 The first subdata Y in (M-1,1) , and the Mth first data Y M The first subdata Y in (M,1)Similarly, in FIG. 4b, the similarity between the second sub-data Y in the first data Y1 is greater than or equal to the first similarity threshold. (1,2) , and the second sub-data Y in the second first data Y2 (2,2) the similarity between the second sub-data Y in the second first data Y2 is greater than or equal to the first similarity threshold; (2,2) , and the second sub-data Y in the third first data Y3 (3,2) The similarity between the (M-1)th first data Y is greater than or equal to the first similarity threshold; M-1 The second subdata Y in (M-1,2) , and the Mth first data Y M The second subdata Y in (M,2) The similarity between the first data and the sub-data is greater than or equal to a first similarity threshold. Other sub-data in the first data may also have the above similarity. For the sake of brevity, details will not be described again. However, this is not a limitation of the present application. For example, any first data may include at least one sub-data, and at least one sub-data does not have the above similarity with other sub-data in the first data.

[0183] The data Y in FIG. 4b is data for h time units, for example data for one time unit, data for two time units, or data for more time units.

[0184] In both Method 1 and Method 2, only an example in which the similarity between subdata at the same position in all first data (e.g., the first subdata in the first data) is greater than or equal to the first similarity threshold is used for explanation. However, this is not a limitation in the present application. For example, the similarity between the second subdata in the first first data and the fourth subdata in the second first data is greater than or equal to the first similarity threshold. In addition, the similarity between subdata may not be limited to the similarity between adjacent first data, but may be the similarity between any two of M first data. For example, the similarity between one subdata in the pth first data and one subdata in the qth first data is greater than or equal to the first similarity threshold. Here, both p and q are positive integers, and p is not equal to q. For example, the similarity between the subdata in the first first data and the subdata in the third first data may be greater than or equal to the first similarity threshold.

[0185] In both Method 1 and Method 2, an example in which each of the different first data has one similar subdata is used for explanation. The number of similar subdata in each of the first data is not limited in this application. For example, the similarity between the first subdata in the qth first data and each of the first and second subdata in the pth first data is greater than or equal to a first similarity threshold. In other words, the first and second subdata in the pth first data share a first dictionary matrix with the first subdata in the qth first data. In another example, the similarity between the first and third subdata in the pth first data and each of the first and second subdata in the qth first data is greater than or equal to a first similarity threshold. In other words, the first and third subdata in the pth first data share a first dictionary matrix with the first and second subdata in the qth first data.

[0186] To improve the compression ratio, in this embodiment of the present application, one sub-data in each of the M first data shares the first dictionary matrix. In other words, the first dictionary matrix includes the characteristics of the M sub-data corresponding to the M first data, respectively. The M sub-data may be sub-data related in the spatial dimension in the above example. For example, the M sub-data may be the nth sub-data Y in the first first data Y1. (1,n) , the n-th sub-data Y in the second first data Y2 (2,n) , ..., and the Mth first data Y M The nth subdata Y in (M,n) may include:

[0187] When the M subdata share a first dictionary matrix, for each of the M subdata, the first device performs dictionary learning on the subdata based on the first dictionary matrix to obtain a first sparse matrix corresponding to the subdata, and performs sparse representation on the subdata based on the first dictionary matrix.

[0188] It should be understood that each of the first data may include at least one subdata correlated with subdata in other first data. When each of the first data includes multiple subdata that are separately correlated with subdata in other first data, the M first data may correspond to multiple groups of subdata, each group of subdata including M subdata, and the M subdata in each group of subdata are correlated with each other. In other words, the M subdata in each group of subdata share one first dictionary matrix. For example, the nth subdata Y in the first first data Y1 may correspond to multiple groups of subdata, each group of subdata including M subdata, and the M subdata in each group of subdata are correlated with each other. In other words, the M subdata in each group of subdata share one first dictionary matrix. For example, (1,n) , the n-th sub-data Y in the second first data Y2 (2,n) , ..., Mth first data Y M The nth subdata Y in (M,n)may be one group of subdata. For ease of explanation, one group of subdata corresponding to M first data is used only as an example for explanation in this embodiment of the present application. Some subdata in the first data may not share a dictionary matrix with other subdata in the first data. Such subdata may be decomposed into a second dictionary matrix and a second sparse matrix based on dictionary learning, and compression and transmission are performed. For ease of understanding in this application, it should be understood that M correlated subdata are represented as one group of subdata, not an actual data set, and the subdata in the first data are not actually grouped.

[0189] The first dictionary matrix may be a pre-configured fixed dictionary matrix, e.g., configured for the first device by the second device, or configured for the second device by the first device; or the first dictionary matrix may be generated or updated by the first device, in which case the first device may transmit the first dictionary matrix to the second device.

[0190] For example, the first dictionary matrix may be obtained through decomposition after the first device performs dictionary learning on the k-th subdata of the M subdata. For example, the first device may decompose the k-th subdata based on the dictionary learning to obtain the first dictionary matrix and a first sparse matrix corresponding to the k-th subdata. To improve the convenience of data processing, the first device may decompose the first subdata of the M subdata to obtain the first dictionary matrix and a first sparse matrix corresponding to the first subdata. The k-th subdata may be obtained as the k-th first data Y k It can be one sub-data in the

[0191] In the above example, for the M-1 subdata other than the k-th subdata among the M subdata, the first device may determine a first sparse matrix corresponding to each of the M-1 subdata based on the first dictionary matrix. Further, M first sparse matrices are obtained.

[0192] It should be understood that a sparse matrix with good sparsity performance has many elements with a value of 0, thereby implementing compression of the first data to some extent. In this embodiment of the present application, the first device may perform data compression on some or all of the M first sparse matrices to further improve compression performance. For example, referring to FIG. 6, the first device may perform data compression on a first sparse matrix X m into first position indication information and an element sequence. The first position indication information indicates the position of an element in the sparse matrix that can represent the corresponding sub-data. The first device sets the value of the first element in the first sparse matrix that cannot represent the corresponding sub-data to a first value based on the indication of the first position indication information. The element sequence includes elements that can represent the corresponding sub-data, i.e., elements that are not set to the first value. The first value may be zero or any other value. Referring to FIG. 6 , in the first sparse matrix, an element in the first row in column x1, an element in the second row in column x2, an element in the second row in column x3, and the like, can represent the corresponding sub-data. In this case, the element sequence includes elements in the sparse matrix that can represent the sub-data, such as an element in the first row in column x1, an element in the second row in column x2, and an element in the second row in column x3. For example, in a sparse matrix, the value of the element in the first row in column x1 is 0.1, the value of the element in the second row in column x2 is -0.5, the value of the element in the second row in column x3 is 0.8... These elements can represent sub-data. The element sequence can be represented as {0.1, -0.5, 0.8...}. However, in a sparse matrix X iThe absolute value of the value of an element at another position other than the element at the above position in x1 is small (including an element whose value is 0 or whose difference between the value and 0 is smaller than a preset capability threshold). An element at such a position cannot represent sub-data. For example, in the first sparse matrix, the value of the element at the second row in column x1 is -0.005, and the value of the element at the first row in column x2 is 0.0002. The first position indication information indicates the position of an element that can represent sub-data in the first sparse matrix. In other words, the first position indication information indicates the position where an element can represent corresponding sub-data in the first sparse matrix. The first position indication information and the element sequence can be used as compressed data of the first sparse matrix.

[0193] The first position indication information may include a tree structure, a bitmap, a position index, and the like, but is not limited thereto. For example, the first position indication information is a bitmap. Referring to FIG. 6 , bits in the bitmap have a one-to-one correspondence with elements in the first sparse matrix, and the bits in the bitmap indicate whether the corresponding elements in the first sparse matrix can represent the corresponding subdata. For example, when a bit in the bitmap is 1, it indicates that the corresponding elements in the first sparse matrix can represent the corresponding subdata; or when a bit in the bitmap is 0, it indicates that the corresponding elements in the first sparse matrix cannot represent the corresponding subdata. In another example, when a bit in the bitmap is 0, it indicates that the corresponding elements in the first sparse matrix can represent the corresponding subdata; or when a bit in the bitmap is 1, it indicates that the corresponding elements in the first sparse matrix cannot represent the corresponding subdata.

[0194] The first position indication information may include at least one position index, and each position index may indicate a position in the first sparse matrix of an element that can represent corresponding sub-data in the first sparse matrix. For example, the position index may indicate that the element in the first sparse matrix that can represent corresponding sub-data is in the second row and the fifth column.

[0195] When the first position indication information includes a tree structure, a basic tree structure corresponding to the first sparse matrix can be constructed, and some or all subnodes in the basic tree structure have a one-to-one correspondence in the first sparse matrix. For example, in the process of constructing the basic tree structure, a multi-level interval division can be performed on a basic two-dimensional plane, and each interval is used as a subnode of the basic tree structure, and the number of the smallest intervals obtained through the division (i.e., the subnodes at the lowest level in the basic tree structure) is greater than or equal to the number of elements in the first sparse matrix. The size of the basic two-dimensional plane is not limited in this application. For example, in the process of constructing the basic tree structure, the basic tree structure can be equally divided into four intervals by using two straight lines perpendicular to each other, and the four intervals respectively correspond to four subnodes of the root node in the basic tree structure. For each of the four intervals, if the interval includes an element capable of representing the corresponding subdata, the value of the subnode is a second value; or if the interval does not include an element capable of representing the corresponding subdata, the value of the subnode is a third value. Furthermore, by using two mutually perpendicular lines, the interval is divided into four subintervals, each of which corresponds to four subnodes at a lower level of the subnode corresponding to the interval in the basic tree structure. Similarly, when a subinterval includes an element capable of representing the corresponding subdata, the value of the subnode corresponding to the subinterval in the basic tree structure is a second value; or, when a subinterval does not include an element capable of representing the corresponding subdata, the value of the subnode corresponding to the subinterval in the basic tree structure is a third value. The rest can be deduced by analogy, and the construction of the tree structure is completed until the minimum interval is obtained. The second value can be 1, and the third value can be 0; or, the second value can be 0, and the third value can be 1; or, the second value and the third value can be any two different values. This is not a limitation in the present application.

[0196] For example, the size of the first sparse matrix of the point cloud data is 3x16. In the process of constructing the basic tree structure, the basic tree structure is equally divided into four intervals by using two mutually perpendicular lines, each corresponding to four subnodes of the root node in the tree structure. Each interval is then equally divided into four subintervals corresponding to four subnodes at a lower level of the subnode corresponding to the interval in the basic tree structure. The remainder can be estimated by analogy, and a 16x16 minimum interval is obtained through four interval divisions (or four recursions), and the basic tree structure of the first sparse matrix is ​​constructed. Some subnodes in the basic tree structure may have a one-to-one correspondence with elements in the first sparse matrix. For example, the subnodes in the basic tree structure corresponding to the 3x16 minimum intervals in the upper left corner of the basic two-dimensional plane each point to a 3x16 element in the first sparse matrix. Optionally, the first position indication information may be a basic tree structure, and the values ​​of nodes in the basic tree structure that correspond to intervals in the basic tree structure that do not include elements in the first sparse matrix are all a third value. Alternatively, the tree structure in the first position indication information may include several subnodes in the basic tree structure, and some of the subnodes are subnodes that correspond to intervals in the basic two-dimensional plane that include elements in the first sparse matrix.

[0197] In the above example, the minimum interval obtained through division includes one element in the first sparse matrix. The present application does not exclude the case where the minimum interval includes multiple elements in the first sparse matrix, or multiple minimum intervals correspond to one element in the first sparse matrix. When the minimum interval includes one element, the location indicated by the first location indication information is more accurate, and therefore, data constructed by the second communication device based on the first location indication information is more accurate. When the minimum interval includes multiple elements, the first location indication information occupies fewer transmission resources, thereby reducing resource overhead.

[0198] The description of the above example of the first location-indicating information may be applicable to the description of the location-indicating information in the following related embodiments, whose implementations are the same or similar. For the sake of brevity, the details will not be described again.

[0199] When the first device is implemented as a chip or a chip system, S220 may be replaced by the first device outputting compressed data of the M first sparse matrices and transmitting the compressed data of the M first sparse matrices by using a terminal device or a network device on which the first device is deployed.

[0200] As described above, the M first data may correspond to multiple groups of sub-data, and each group of sub-data is represented by M corresponding first sparse matrices based on the same first dictionary matrix. After performing protocol encapsulation, the first device may transmit compressed data of the first sparse matrices corresponding to all of the multiple groups of sub-data together; or after performing encapsulation, the first device may transmit compressed data of the M first sparse matrices corresponding to one group of sub-data together; or after performing encapsulation, the first device may transmit compressed data of each first sparse matrix separately.

[0201] The first device may further transmit the first dictionary matrix to the second device. Optionally, when the first device is implemented as a chip or a chip system, the first device outputs the first dictionary matrix, and a terminal device or a network device on which the first device is deployed transmits the first dictionary matrix.

[0202] After receiving the compressed data of the M first sparse matrices sent by the first device, the second device may decompress the compressed data of the M first sparse matrices to obtain decompressed data. Decompressing the M first sparse matrices may alternatively be expressed as restoring the M first sparse matrices or constructing the M first sparse matrices, and it should be noted that for ease of explanation, the meaning expressed after the word order is changed is still consistent. For example, data decompression is performed on the M first sparse matrices, or data construction is performed on the M first sparse matrices.

[0203] The second device may construct M first sparse matrices based on the first dictionary matrix and the compressed data.

[0204] In some embodiments, the first device may perform a compression process on the compressed data of the first dictionary matrix and / or the M first sparse matrices. The compression process includes, but is not limited to, quantization (e.g., scalar quantization or vector quantization) and / or entropy coding. Correspondingly, the second device needs to receive the compressed data of the first dictionary matrix and / or the M first sparse matrices obtained through the compression process and perform a corresponding decompression process on the compressed data of the first dictionary matrix and / or the M first sparse matrices obtained through the compression process.

[0205] Therefore, in this embodiment of the present application, based on one first dictionary matrix, dictionary learning is performed on M sub-data corresponding to M first data respectively, to obtain a sparse representation of each of the sub-data, thereby implementing effective and reliable data compression and transmission in transmission scenarios involving large data volumes.

[0206] To further improve the compression ratio, in this embodiment of the present application, joint compression may be performed on the M first sparse matrices. The following two implementations are used as examples for illustration:

[0207] Implementation 1: Perform joint compression on M first sparse matrices through low-rank approximation.

[0208] In implementation 1, the first device may determine a first matrix based on the M first sparse matrices, and perform low-rank approximation on the first matrix to obtain first compressed data of the M first sparse matrices. The manner of obtaining the M first sparse matrices is described in the above example. The details will not be described again here.

[0209] In a first example of Implementation 1, the first device may combine M first sparse matrices to obtain a first matrix. For example, M) are stacked by row, and the first matrix

number

number

[0210] In a second example of implementation 1, the first device may first perform data compression on at least one of the M first sparse matrices, and combine the M first sparse matrices obtained through the data compression to obtain a first matrix.

[0211] In the second example, the first sparse matrix X m is used as an example, and the first sparse matrix obtained through data compression is

number

number

number

[0212] When performing the above data compression on the plurality of first sparse matrices in the M number of first sparse matrices, the first device may generate the same first position indication information (i.e., B1=B2==B M ), the first device may perform data compression on a plurality of first sparse matrices. Specifically, the first device sets the value of a first element in each first sparse matrix to a first value based on the same first position indication information. In this case, the first position indication information may be determined based on one of the M first sparse matrices. The one sparse matrix may be a first sparse matrix corresponding to a first subdata in the M first sparse matrices, or may be any first sparse matrix in the M first sparse matrices. For example, the first device determines first position indication information and an element sequence corresponding to one first sparse matrix. Furthermore, when performing data compression on M−1 sparse matrices other than the one first sparse matrix, the first device may select, based on the first position indication information, an element in each first sparse matrix that can represent the corresponding subdata, i.e., set the value of an element that cannot represent the corresponding subdata to a first value, to obtain an element sequence for each first sparse matrix.

[0213] When performing the above data compression on a plurality of first sparse matrices in the M first sparse matrices, the first device may determine first position indication information corresponding to each of the plurality of first sparse matrices, and then, for each of the plurality of first sparse matrices, perform data compression on the first sparse matrix based on the first position indication information corresponding to the first sparse matrix.

[0214] Optionally, the first device may determine first position indication information based on a result of comparing each element in the first sparse matrix with the capability threshold. For example, a bit 1 in the first position indication information indicates the position of an element whose absolute value is greater than the capability threshold, i.e., an element capable of representing the corresponding subdata, and a bit 0 in the first position indication information indicates the position of an element whose absolute value is less than the capability threshold, i.e., an element unable to represent the corresponding subdata. An element whose absolute value is equal to the capability threshold may be determined as an element capable of representing the corresponding subdata or an element unable to represent the corresponding subdata. This is not limited in the present application. The first device may further determine the first position indication information based on the first W elements with the largest absolute values ​​in the first sparse matrix. For example, the first device may determine the first W elements with the largest absolute values ​​in the first sparse matrix as elements capable of representing the corresponding subdata, and determine elements other than the W elements in the first sparse matrix as elements unable to represent the corresponding subdata. Furthermore, bit 1 in the first position indication information indicates the position of an element that can represent the corresponding sub-data, and bit 0 indicates the position of an element that cannot represent the corresponding sub-data. It should be understood that the bit value indicating the position of an element that has or cannot represent the corresponding sub-data is not limited in this application. For example, bit 1 may further indicate the position of an element that cannot represent the corresponding sub-data, and bit 0 may indicate the position of an element that can represent the corresponding sub-data.

[0215] In the second example, the first device may obtain a first matrix by combining a sequence of elements of a first sparse matrix obtained through data compression, for example, a first sparse matrix obtained after element selection based on expressive power. For example, M first sparse matrices (

number

number

number

number

[0216] In both the first and second examples of Implementation 1, the first matrix is ​​a high-dimensional matrix, and direct transmission of the first matrix would cause high communication overhead. Therefore, the first device may perform singular value decomposition on the first matrix to obtain K feature values ​​(also referred to as singular values) and a feature vector (also referred to as a singular value vector) respectively corresponding to the K feature values. The K feature values ​​and the feature vector respectively corresponding to the K feature values ​​may represent the first matrix. For example, the first matrix

number

number

number

number

number

[0217] As described above, the M first data may correspond to multiple groups of sub-data, each group of sub-data being represented by using M corresponding first sparse matrices based on the same first dictionary matrix, and one first compressed data is formed based on the M first sparse matrices representing corresponding sub-data based on the same first dictionary matrix. After performing protocol encapsulation, the first device may transmit the first compressed data corresponding to each of the multiple groups of sub-data together, or after performing protocol encapsulation, the first device may transmit each of the first compressed data separately.

[0218] In some embodiments, the first device transmits compressed data of M first data in an incremental manner. In this case, the first compressed data may be compressed data of the M first data transmitted in an initial transmission process. In the incremental transmission process, the first device may transmit second compressed data of the M first sparse matrices to the second device. The second compressed data may include K1 characteristic values ​​other than the K0 characteristic values ​​among the K characteristic values ​​obtained through singular value decomposition, and characteristic vectors corresponding to the K1 characteristic values, respectively. The K0 characteristic values ​​may be the first K0 characteristic values ​​among the K characteristic values, and the K1 characteristic values ​​may be the first K1 characteristic values ​​other than the first K0 characteristic values ​​among the K characteristic values, for example, the (K0+1)th characteristic value to the (K0+K1)th characteristic value. Correspondingly, the characteristic vectors corresponding to the K1 characteristic values ​​may include the (K0+1)th to (K0+K1)th column vectors of the left singular matrix and the (K0+1)th to (K0+K1)th row vectors of the right singular matrix.

[0219] It may be understood that the first device may incrementally transmit one or more of the sub-data of the plurality of groups. When incrementally transmitting the sub-data of at least two groups, the first device may transmit second compressed data corresponding to the sub-data of the plurality of groups together after performing protocol encapsulation, or the first device may transmit each of the second compressed data separately after performing protocol encapsulation.

[0220] It should be further understood that the first device may perform one or more incremental transmissions. For example, during the second incremental transmission, the first device may transmit third compressed data to the second device. The third compressed data may include K2 characteristic values ​​other than the K0+K1 characteristic values ​​among the K characteristic values, and characteristic vectors corresponding to the K2 characteristic values, respectively.

[0221] When the first device is implemented as a chip or a chip system, the first device may output the second compressed data and transmit the second compressed data by using a terminal device or a network device in which the first device is deployed and by using a transceiver.

[0222] Implementation 2: A residual-based scheme performs joint compression on the M first sparse matrices.

[0223] In Implementation 2, the first device may sequentially transmit compressed data of all M first sparse matrices, and each time compressed data of a first sparse matrix is ​​transmitted, currently transmitted compressed data of the first sparse matrix may be determined based on the compressed data of the previously transmitted first sparse matrix. For ease of explanation, the jth first sparse matrix among the M first sparse matrices is used as an example. The first device determines first residual information based on information about the jth first sparse matrix and information about the ith first sparse matrix, and transmits the first residual information to the second device by using the first residual information as compressed data of the jth first sparse matrix, where i is smaller than j and both i and j are positive integers. The information about the ith first sparse matrix may be obtained by decompressing the compressed data of the ith first sparse matrix.

[0224] In some embodiments, any one of the M first sparse matrices is used as a base sparse matrix. The first device may use the first position indication information and the element sequence of the base sparse matrix as compressed data for the base sparse matrix, and for each of the M-1 first sparse matrices other than the base sparse matrix, the first device performs compression and transmission for each first sparse matrix based on the compression and transmission method for the jth first sparse matrix. The base sparse matrix may be the first first sparse matrix to be compressed and transmitted among the M first sparse matrices.

[0225] In some other embodiments, the M first sparse matrices are grouped into multiple sparse matrix groups, and any first sparse matrix in each sparse matrix group is used as a base sparse matrix in the group. The first device may use the first position indication information and the element sequence of the base sparse matrix as compressed data for the base sparse matrix, and for the first sparse matrices other than the base sparse matrix in the group, the first device performs compression and transmission for each first sparse matrix in the group based on the compression and transmission method used for the jth first sparse matrix. The base sparse matrix may be the first sparse matrix in the sparse matrix group that is compressed and transmitted for the first time.

[0226] For example, the first device may determine a residual matrix based on information about the jth first sparse matrix and information about the ith first sparse matrix, and then determine first residual information based on the residual matrix. The first residual information may include a first residual element sequence, and the first residual element sequence includes residual elements in the residual matrix whose absolute values ​​are greater than or equal to a first residual threshold, and elements whose absolute values ​​are greater than or equal to the first residual threshold and can represent a residual between the information about the jth first sparse matrix and the information about the ith first sparse matrix. The second device may construct the jth first sparse matrix based on the first residual element sequence and the information about the ith first sparse matrix. Optionally, when a residual matrix between a matrix obtained by decompressing the compressed data of the jth first sparse matrix and the ith first sparse matrix is ​​obtained, a difference between values ​​of elements at each corresponding position in the two matrices may be calculated to obtain the residual matrix.

[0227] In the above example, the first residual information may further include second position indication information. The second position indication information indicates the position of a residual element in the residual matrix between the information about the jth first sparse matrix and the information about the ith first sparse matrix, the absolute value of which is greater than a first residual threshold, i.e., the position of an element in the residual matrix that can represent a residual between the information about the jth first sparse matrix and the information about the ith first sparse matrix. It should be noted that the second position indication information may further indicate the position of a residual element in the residual matrix whose absolute value is greater than the first residual threshold, i.e., the position of an element in the residual matrix that cannot represent a residual between the information about the jth first sparse matrix and the information about the ith first sparse matrix. An element in the residual matrix whose absolute value is equal to the first residual threshold may be considered to be capable of representing a residual or not capable of representing a residual. This is not a limitation in the present application. Bits indicating two residual elements that can or cannot represent a residual between the information about the jth first sparse matrix and the information about the ith first sparse matrix may be different.

[0228] For example, the first device may obtain a residual matrix based on the difference between the jth first sparse matrix and the restored ith first sparse matrix; or the first device may obtain the jth first sparse matrix (e.g.,

number

number

number

number

number

[0229] To further improve the compression rate of the M first sparse matrices, in some embodiments of Implementation 2, the first device may determine a similarity between the jth first sparse matrix and the ith first sparse matrix. When the similarity between the jth first sparse matrix and the ith first sparse matrix is ​​less than a second similarity threshold, the first device transmits first residual information to the second device; when the similarity between the jth first sparse matrix and the ith first sparse matrix is ​​greater than or equal to the second similarity threshold, the first device does not transmit compressed data of the jth first sparse matrix; when the similarity between the jth first sparse matrix and the ith first sparse matrix is ​​equal to the second similarity threshold, the compressed data of the jth first sparse matrix may or may not be transmitted. This is not a limitation of the present application.

[0230] When the first device does not transmit the compressed data of the jth first sparse matrix, the second device uses the compressed data of the ith first sparse matrix as the compressed data of the jth first sparse matrix. In other words, the second device uses the constructed ith first sparse matrix as the jth first sparse matrix.

[0231] The similarity between the jth first sparse matrix and the ith first sparse matrix can be determined based on the residual matrix between the jth first sparse matrix and the ith first sparse matrix. For example, the similarity between the jth first sparse matrix and the ith first sparse matrix can be determined based on the modulus value of the residual matrix (e.g.,

number

number

number

number

number

number

[0232] It should be understood that before transmitting the compressed data of the jth first sparse matrix to the second device, the first device may transmit compressed data of at least one first sparse matrix to the second device. In this case, the ith first sparse matrix may be any one of the at least one first sparse matrix. To facilitate compression, transmission, and data recovery, the ith first sparse matrix may be the first sparse matrix transmitted in one transmission closest to the jth first sparse matrix among the at least one first sparse matrix. When the ith first sparse matrix is ​​the first first sparse matrix, in Implementation 2, the compressed data of the ith first sparse matrix includes first position indication information and an element sequence of the ith first sparse matrix; when the ith first sparse matrix is ​​not the first first sparse matrix, in Implementation 2, the compressed data of the ith first sparse matrix includes second position indication information and a first residual element sequence of the ith first sparse matrix.

[0233] As described above, the M first data may correspond to multiple groups of sub-data, and each group of sub-data is represented by using M corresponding first sparse matrices based on the same first dictionary matrix. Compressed information of the first sparse matrices corresponding to the data in each group may be transmitted together after protocol encapsulation is performed. For example, the nth sub-data Y in the first data Y is (1,n) , the n-th sub-data Y in the second first data Y2 (2,n) , ..., and the Mth first data Y M The nth subdata Y in (M,n) may be transmitted together after protocol encapsulation is performed, or the compressed data of the first sparse matrix corresponding to the data in each group may be transmitted separately after protocol encapsulation is performed.

[0234] When the first device is implemented as a chip or a chip system, the first device may output compressed data of M first sparse matrices and transmit second compressed data by using a terminal device or a network device on which the first device is deployed and by using a transceiver.

[0235] In some embodiments, the first device transmits compressed data of the M first data in an incremental manner. In this case, the first residual information transmitted in the above embodiments may be compressed data of the jth first sparse matrix transmitted in the initial transmission process. In the incremental transmission process, the first device may transmit second residual information of the jth first sparse matrix to the second device. The second residual information is also determined based on the information about the jth first sparse matrix and the information about the ith first sparse matrix. A difference between the second residual information and the first residual information is that third position indication information in the second residual information indicates a residual element in the residual matrix between the information about the jth first sparse matrix and the information about the ith first sparse matrix that is smaller than or equal to the first residual threshold and larger than or equal to the second residual threshold, and the second residual element sequence in the second residual information includes residual elements in the residual matrix that are smaller than or equal to the first residual threshold and larger than or equal to the second residual threshold. The second residual information transmitted in the increment process can complement the first residual information and enrich the residual information, so that the second device can more accurately construct the jth first sparse matrix based on the first residual information and the second residual information.

[0236] Furthermore, the first device may transmit third residual information when performing incremental transmission again in the incremental transmission process. The third residual information is also determined based on the information about the jth first sparse matrix and the information about the ith first sparse matrix, and the fourth position indication information in the third residual information indicates a residual element in the residual matrix between the information about the jth first sparse matrix and the information about the ith first sparse matrix that is smaller than or equal to the second residual threshold and larger than or equal to the third residual threshold, thereby complementing the first residual information and the second residual information. Naturally, the first device may transmit more residual information. The number of incremental transmissions in the incremental transmission process is not limited in this application.

[0237] It should be noted that the first device may incrementally transmit a first sparse matrix corresponding to one or more of the M subdata, and / or the first device may incrementally transmit one or more of a plurality of groups of subdata corresponding to the M first data.

[0238] In Implementation 1 and Implementation 2, the first device may perform a compression process on the compressed data of the M first sparse matrices, such as performing at least one of scalar quantization, vector quantization, and entropy coding, and transmit first residual information obtained through the compression process to the second device, which then performs a corresponding decompression process on the compressed data. The compressed data of the M first sparse matrices may include the first compressed data and / or the second compressed data in Implementation 1, or may include the first residual information and / or the second residual information in Implementation 2.

[0239] All thresholds such as the capability threshold, the similarity threshold, and the residual threshold (including the first to fourth residual thresholds) in Implementation 1 and Implementation 2 may be pre-set or pre-configured. Pre-setting a threshold may mean, for example, that the threshold is defined in a protocol or pre-stored in a device, and pre-configuring a threshold may mean, for example, that the threshold is pre-configured for a terminal device by a network device.

[0240] 7 is a schematic interaction flowchart of another data compression and transmission method according to an embodiment of the present application. As shown in FIG. 7, a method 300 includes some or all of the processes in S310 to S390.

[0241] S310: The first device sends a compress and transmit request to the second device, where the compress and transmit request carries the data type of the data to be transmitted. Correspondingly, the second device receives the compress and transmit request sent by the first device.

[0242] S320: The second device sends first configuration information to the first device, where the first configuration information indicates first time-frequency resources of the M first data. Correspondingly, the first device receives the first configuration information and / or the first indication information sent by the second device.

[0243] S330: The first device transmits compressed data of the M first sparse matrices and first indication information on a first time-frequency resource, where the first indication information indicates a first compression parameter. Correspondingly, the second device receives the compressed data of the M first sparse matrices and the first indication information transmitted by the first device on the first time-frequency resource.

[0244] S340: The second device constructs M pieces of first data based on the compressed data of the M first sparse matrices and the first instruction information.

[0245] S350: The second device transmits second configuration information to the first device, where the second configuration information indicates second time-frequency resources of the compressed data of the M first sparse matrices. Correspondingly, the first device receives the second configuration information transmitted by the second device.

[0246] S360: The first device transmits the compressed data of the M first sparse matrices and the second indication information on the second time-frequency resource. Correspondingly, the second device receives the compressed data of the M first sparse matrices and the second indication information transmitted by the first device on the second time-frequency resource.

[0247] S370: The second device updates the M constructed first data.

[0248] It should be noted that the initial transmission process is the basis for the incremental transmission process. When incremental transmission is not present, the initial transmission process includes a complete compression and transmission process. When method 300 includes S350-S370, S310-S340 are the initial transmission process, and S350-S370 are the incremental transmission process. When method 300 does not include S350 and S370, S310-S340 are the compression and transmission process.

[0249] When steps S310 to S340 are implemented as an initial transmission process, the first time-frequency resource is the time-frequency resource occupied in the initial transmission process, and the compressed data of the M first data transmitted on the first time-frequency resource in the initial transmission process is, for example, the first compressed data in Implementation 1 and the first residual information in Implementation 2, and may further include compressed data of the first dictionary matrix. The first compression parameter is the compression parameter in the initial transmission process. The second time-frequency resource is the time-frequency resource occupied in the incremental transmission process. The compressed data of the M first sparse matrices is compressed data transmitted on the second time-frequency resource in the incremental transmission process, for example, the second compressed data in Implementation 1 and the second residual information in Implementation 2.

[0250] Optionally, the data type includes point cloud data or AI model data.

[0251] Optionally, the compress and transmit request may be a status report (SR) or a buffer status report (BSR). For example, the first device may send the compress and transmit request by using an SR to indicate a data type, or the first communication device may send the compress and transmit request by using a BSR to report the size of the data to be transmitted (e.g., the M first sparse matrices and / or the first dictionary matrix).

[0252] In S310, the first device sends a compression and transmission request to the second device to request that compression and transmission be performed on the M first sparse matrices. The compression and transmission request may include a data type of data to be transmitted and / or a size of the data to be transmitted.

[0253] The second device may determine first time-frequency resources for transmitting the compressed data of the M first sparse matrices based on the data type of the data to be transmitted and / or the size of the data to be transmitted. Optionally, when the first data is large and the compressed data of the M first sparse matrices need to be transmitted in an incremental transmission manner, the second device may first determine first time-frequency resources for the compressed data of the M first sparse matrices in an initial transmission process.

[0254] Before S330, in this embodiment, the compressed data of the M first sparse matrices may further be determined based on the method in any one of the above embodiments, and for the sake of brevity, the details will not be described again.

[0255] According to any one of the above embodiments, the first device may determine a first compression parameter based on the first time-frequency resource, and then perform data compression based on the first compression parameter. The first compression parameter may include at least one of the following: (1) a capacity threshold, where the capacity threshold is used to determine whether an element in the first sparse matrix can represent one corresponding sub-data in the first data; (2) a first residual threshold, where the first residual threshold is used to determine a first residual element sequence, where the first residual element sequence represents a residual matrix between information about the jth first sparse matrix and information about the ith first sparse matrix, where i is less than j and both i and j are positive integers; and (3) compression processing parameters, where the compression processing parameters include at least one of a quantization precision, a quantization codebook, and an encoding scheme. The compression processing parameters may indicate the encoding scheme, quantization precision, quantization codebook, and the like of each of the first dictionary matrix, the first compressed data, and the first residual information. Optionally, the coding scheme may include scalar quantization, vector quantization, entropy coding, and the like.

[0256] Optionally, in the scalar quantization scheme, the compression processing parameters may further include a quantization range of the scalar quantization; in the vector quantization scheme, the compression processing parameters may further include a quantization codebook corresponding to the vector quantization.

[0257] In S330, the first instruction information and the first compressed data may be transmitted together on the first time-frequency resource. However, this is not a limitation in the present application. For example, the first instruction information may be transmitted separately, and the order of transmitting the first instruction information and the first compressed data by the first device is not limited. Alternatively, the first instruction information may be transmitted by the second device to the first device to instruct the first device to compress data, or the content indicated by the first instruction information may be agreed upon in a protocol.

[0258] The first indication information indicates at least one of the following: (1) the number K0 of low-rank approximation characteristics: For example, K0 may be equal to 4 or 8 based on Scheme 2 in which data division is performed in the spatial dimension. However, this is not limited to this application. The value of K0 may be adaptively adjusted based on an application scenario or a communication service; (2) the number M of first data: The number M of first data may be adaptively adjusted based on an application scenario or a communication service. Optionally, the number M of first data may be related to K0. For example, based on Scheme 2 in which data division is performed in the temporal dimension, when K0 is equal to 4, M may be equal to 2, and when K0 is equal to 8, M may be equal to 4. Based on Scheme 1 in which data division is performed in the temporal dimension, when the data to be transmitted is AI data, the number M of first data may be equal to 10; when the data to be transmitted is point cloud data, the number M of first data is related to a scanning frequency at which the point cloud data is collected. For example, if the rotational scanning frequency of the lidar is normally 20 Hz, the number M of first data may be equal to 5. (3) a capability threshold; (4) a first residual threshold; (5) a proportion of elements in one of the M first sparse matrices that can represent one corresponding sub-data in the first data; (6) data loss of the first compressed data for the M first sparse matrices; (7) compression process parameters; and (8) whether to transmit first residual information.

[0259] (1) the number of characteristics of the low-rank approximation K0; (2) the capacity threshold; (4) the first residual threshold; (5) the proportion of elements in the first sparse matrix that can represent one corresponding subdata in the first data; and (6) the data loss of the first compressed data can all be used by the second device to determine whether the M first data (or the M first sparse matrices) need to be compressed and transmitted in an incremental transmission manner. (4) the first residual threshold and (7) the compression process parameters can instruct the first device to perform data compression and instruct the second device to decompress the first compressed data. (1) the number of characteristics of the low-rank approximation K0; (2) the number of first data M; (3) the capacity threshold; (4) the first residual threshold; and (8) whether to transmit the first residual information can all be used by the first device to perform data compression and by the second device to perform data construction.

[0260] For example, in implementation 1, the first indication information may include one or more of the fields shown in Table 1. Table 1 [Table 1]

[0261] The compression status is used by the second device to determine whether the first data needs to be compressed and transmitted in an incremental transmission manner. The compression status field may include at least one of: (1) a number K0 of characteristics of the low-rank approximation; (2) a capacity threshold; (4) a first residual threshold; (5) a proportion of elements in the first sparse matrix that can represent one corresponding subdata in the first data; and (6) a data loss of the first compressed data.

[0262] Optionally, in the first instruction information shown in Table 1, the compression instruction field of the first dictionary matrix may indicate the compression precision of the first dictionary matrix. For example, if one group of subdata in the M first data shares the first dictionary matrix, it may be instructed that high-precision (e.g., 32-bit) quantization is performed on the first dictionary matrix by using a quantizer. The compression instruction of the K0 characteristic values ​​is the same as the compression instruction of the first dictionary matrix. Details will not be described again. Additionally / alternatively, the compression instruction field of the first dictionary matrix may indicate a compression method of the first dictionary matrix, such as quantization or entropy coding.

[0263] Optionally, in the first indication information shown in Table 1, the compression indications of the feature vectors corresponding to the K0 feature value fields indicate compression methods of the feature vectors, such as quantization, entropy coding, dictionary learning, etc. Furthermore, at least one of a quantization precision, a quantization codebook, and a coding method may be indicated.

[0264] Optionally, the data partitioning mode field may indicate that data partitioning is to be performed in the time dimension in scheme 1, or in scheme 2 that data partitioning is to be performed in the space dimension.

[0265] Optionally, the data division dimension field may include an indication of the number M of first data, an indication of the number of sub-data in the first data, or the like.

[0266] In some embodiments, the compression instruction field of the first dictionary matrix in Table 1, the compression instruction field of the K0 feature values, and the compression instruction field of the feature vector corresponding to the K0 feature values ​​are all required fields, and the data partitioning mode field, the data partitioning dimension field, and the compression status field are all optional fields.

[0267] For example, in implementation 2, the first indication information may include one or more of the fields shown in Table 2. Table 2 [Table 2]

[0268] The data partition mode field, data partition dimension field, and compression status field are all the same as in Table 1. The details will not be repeated here.

[0269] Optionally, any one of the compression instruction fields, such as the compression instruction field of the first dictionary matrix, the compression instruction field of the first position indication information, the compression instruction field of the second position indication information, and the compression instruction field of the first residual element sequence, may each indicate a compression method and / or a compression precision of the corresponding data. For example, the compression instruction field of the first dictionary matrix indicates that 32-bit high-precision quantization is to be performed on the first dictionary matrix. In another example, the compression instruction field of the first position indication information indicates that the first position indication information is to be compressed using the Lempel-Ziv 77 (LZ77) or Lempel-Ziv-Markov Chain Algorithm (LZMA) method, and the compression instruction field of the second position indication information indicates that the second position indication information is to be compressed using the LZ77 or LZMA method. LZ77 is a dictionary-based sliding window lossless compression algorithm invented by Lempel-Ziv in 1977. LZMA is a compression algorithm improved based on LZ77, characterized by high compression ratio, high decompression speed, and low memory consumption. In another example, the compression instruction field of the element sequence indicates the quantization precision and / or quantization range when scalar quantization is performed on the element sequence, or the compression instruction field of the element sequence indicates the quantization precision and / or quantization codebook when vector quantization is performed on the element sequence, or the compression instruction field of the element sequence indicates the coding method when entropy coding is performed on the element sequence. The first residual element sequence is similar and will not be described again.

[0270] The above compression methods are merely examples and are not limiting. The compression method of the first position indication information may further include any other compression algorithm. Examples are not listed.

[0271] The first position indication information and element sequence are determined based on the base sparse matrix, and the second position indication information and first residual element sequence are determined based on information about the jth first sparse matrix and information about the ith first sparse matrix.

[0272] Optionally, the field of whether to transmit first residual information may indicate whether to transmit first residual information corresponding to the jth first sparse matrix by using 1 bit. For example, when the field of whether to transmit first residual information is 0, it indicates not to transmit first residual information corresponding to the jth first sparse matrix; or when the field of whether to transmit first residual information is 1, it indicates to transmit first residual information corresponding to the ith first sparse matrix. Of course, the value of the field of whether to transmit first residual information is not limited in the present application.

[0273] In a scenario with a high delay requirement, the first instruction information may instruct to perform joint compression on the M first sparse matrices in the above residual-based scheme to reduce the impact of delay on communication services. In a scenario with a low delay requirement, the first instruction information may instruct to perform joint compression on the M first sparse matrices in the above low-rank approximation-based scheme.

[0274] In S340, the second device decompresses the compressed data of the M first sparse matrices based on the instructions of the first instruction information, constructs M sub-data corresponding to the M first sparse matrices, and performs data construction on other sub-data in the M first data in a similar manner to obtain M first data.

[0275] In S350, the second device may determine whether incremental transmission needs to be performed based on the construction quality of the M first data constructed in S340 and / or the compression status of the compressed data of the M first sparse matrices, and may transmit second configuration information to the first device when incremental transmission needs to be performed. For example, when the construction quality of the M first data (or the M subdata) is low, the second device generates and transmits the second configuration information. In another example, when the compression status field indicates that data loss of the first compressed data for the M first sparse matrices exceeds a threshold, the second device transmits the second configuration information. The compression status may be indicated by the compression status field.

[0276] According to any one of the above embodiments, the first device may calculate the total number of bits that can be transmitted based on the second time-frequency resource; determine a second compression parameter; and perform data compression based on the second compression parameter. The second compression parameter includes a second residual threshold, so that the first device determines the residual element to be transmitted in the incremental transmission process by referring to the second residual threshold. Correspondingly, the second device determines the residual element to be transmitted by the first device in the incremental transmission process by referring to the second residual threshold.

[0277] In S360, the second instruction information may be transmitted on the second time-frequency resource together with the compressed data of the M first sparse matrices (e.g., the first compressed data or the second residual information). However, this is not a limitation in the present application. For example, the second instruction information may be transmitted separately, and the order of transmitting the second instruction information and the compressed data by the first device is not limited. Alternatively, the second instruction information may be transmitted by the second device to the first device and may instruct the first device to perform data compression in the incremental transmission process; or the content indicated by the second instruction information may be agreed upon in a protocol.

[0278] The second instruction information may instruct to perform incremental transmission through low-rank approximation. In this case, the second instruction information may instruct the number K1 of characteristics of the low-rank approximation; or the second instruction information may instruct to perform incremental transmission based on residual. In this case, the second instruction information may instruct a second residual threshold, for example, a second residual threshold.

[0279] For example, in implementation 1, the second instruction information may include one or more of the fields shown in Table 3. Table 3 [Table 3]

[0280] Optionally, the incremental transmission mode indication field may indicate whether incremental transmission is performed through low-rank approximation or in a residual-based manner.

[0281] Optionally, the position indication field may indicate a first data item that needs to be incrementally transmitted among the M pieces of first data, or a sub-data item that needs to be incrementally transmitted among the first data.

[0282] For the compression instruction field for the K1 characteristic values, please refer to the description of the compression instruction field for the K0 characteristic values ​​in the example above. For the compression instruction field for the characteristic vector corresponding to the K1 characteristic values, please refer to the description of the compression instruction field for the characteristic vector corresponding to the K0 characteristic values ​​in the example above. For the sake of brevity, the details will not be repeated.

[0283] For example, in implementation 2, the second instruction information may include one or more of the fields shown in Table 4. Table 4 [Table 4]

[0284] For the increment transmission mode indication field and the position indication field, see the description in Table 3. For the compression indication field of the third position indication information, see the description of the compression indication field of the second position indication information in the above example. For the compression indication field of the second residual element sequence, see the description of the compression indication field of the first residual element sequence in the above example.

[0285] In S370, the second device may update the constructed M pieces of first data based on the second compressed data of the M first sparse matrices or the complementation of the second residual information for the constructed M pieces of sub-data.

[0286] 7 is described as an example using only one incremental transmission. The number of incremental transmissions is not limited in this embodiment of the present application. The first device and the second device may further perform incremental transmissions based on S350 to S370 or in a similar manner.

[0287] FIG. 8 is a schematic block diagram of a communication device according to an embodiment of the present application. The communication device 400 may be a terminal or a network device, may be a device in a terminal device or a network device, or may be a device usable with a terminal device or a network device. In a possible implementation, the communication device 400 may include modules or units that correspond one-to-one to the methods / operations / steps / actions performed by the first device or the second device in the above method embodiments, and the units may be implemented by hardware circuits, software, or a combination of hardware circuits and software. In a possible implementation, as shown in FIG. 8, the device 400 may include a processing module 410 and a transceiving module 420.

[0288] Optionally, the communication device 400 may correspond to the first device in the above method embodiments.

[0289] When the communication device 400 is configured to perform the method on the first device side, the processing module 410 may be configured to obtain M first data, where one sub-data in the first data corresponds to one first sparse matrix, the first sparse matrix represents one corresponding sub-data in the first data based on a first dictionary matrix, the first dictionary matrix includes characteristics of the M sub-data respectively corresponding to the M first data, and the transceiver module 420 may be configured to output compressed data of the M first sparse matrices, where M is an integer greater than 1.

[0290] It should be understood that the specific processes performed by the modules are described in detail in the above method embodiments, and for the sake of brevity, the details will not be described again here.

[0291] Optionally, the communication device 400 may correspond to the second device in the above method embodiments.

[0292] When the communication device 400 is configured to perform the method on the second device side, the transceiver module 420 may be configured to receive compressed data of M first sparse matrices, where the first sparse matrices represent corresponding sub-data in the first data based on a first dictionary matrix, the first dictionary matrix includes characteristics of the M sub-data corresponding to the M first data respectively, and one sub-data in the first data corresponds to one first sparse matrix; and the processing module 410 may be configured to output decompression information based on the compressed data.

[0293] It should be understood that the specific processes performed by the modules are described in detail in the above method embodiments, and for the sake of brevity, the details will not be described again here.

[0294] The transceiver module 420 in the communication device 400 may be implemented by using a transceiver and may correspond, for example, to the transceiver 520 in the communication device 500 shown in Figure 9. The processing module 410 in the communication device 400 may be implemented by using at least one processor and may correspond, for example, to the processor 510 in the communication device 500 shown in Figure 9.

[0295] When the communication device 400 is a chip or a chip system configured in a communication device (e.g., a terminal device or a network device), the transceiver module 420 in the communication device 400 may be implemented by using an input / output interface, circuit, or the like, and the processing module 410 in the communication device 400 may be implemented by using a processor, microprocessor, integrated circuit, or the like integrated on the chip or chip system.

[0296] 9 is another schematic block diagram of a communication device according to an embodiment of the present application. As shown in FIG. 9, the communication device 500 may include a processor 510. The processor 510 may be configured to execute the method performed by the first device or the second device in the above-described method embodiments.

[0297] In some possible implementations, the communications device 500 may include a transceiver 520. The transceiver 520 and the processor 510 may communicate with each other through an internal connection path. The processor 510 may control the transceiver 520 to transmit signals and / or receive signals.

[0298] In some possible implementations, the communication device 500 may include a memory 530. The memory 530 and the processor 510 may communicate with each other through an internal connection path. The memory 530 and the processor 510 may be integrated together or may be located separately. The memory 530 may alternatively be a memory external to the device. The memory 530 is configured to store instructions, and the processor 510 is configured to execute the instructions stored in the memory 530 to perform the method in the above method embodiments.

[0299] It should be understood that the communication device 500 may correspond to a terminal device or a network device in the above method embodiments and may be configured to perform steps and / or procedures performed by the first device or the second device in the above method embodiments. Optionally, the memory 530 may include a read-only memory and a random access memory to provide instructions and data to the processor. A portion of the memory may further include a non-volatile random access memory. The memory 530 may be a separate component or may be integrated into the processor 510. The processor 510 may be configured to execute instructions stored in the memory 530. When the processor 510 executes the instructions stored in the memory, the processor 510 is configured to perform steps and / or procedures corresponding to the terminal device or the network device in the above method embodiments.

[0300] Optionally, the communication device 500 is the first device in the above embodiments.

[0301] Optionally, the communication device 500 is the second device in the above embodiments.

[0302] The transceiver 520 may include a transmitter and a receiver. The transceiver 520 may further include one or more antennas. The processor 510, the memory 530, and the transceiver 520 may be integrated devices on different chips. For example, the processor 510 and the memory 530 may be integrated on a baseband chip, while the transceiver 520 may be integrated on a radio frequency chip. Alternatively, the processor 510, the memory 530, and the transceiver 520 may be integrated devices on the same chip. This is not a limitation of the present application.

[0303] Optionally, the communication apparatus 500 is a component, for example a chip or a chip system, configured in a terminal device.

[0304] Optionally, the communication apparatus 500 is a component configured in a network device, for example, a chip or a chip system.

[0305] The transceiver 520 may alternatively be a communication interface, such as an input / output interface or circuit. The transceiver 520, the processor 510, and the memory 530 may be integrated on the same chip, for example, on a baseband chip.

[0306] The present application further provides a processing device including at least one processor configured to execute a computer program or logic circuit, such that the processing device performs the method performed by the first device or the second device in the method embodiments described above. The processing device may further include a memory configured to store the computer program.

[0307] An embodiment of the present application further provides a processing device including a processor and an input / output interface. The input / output interface is coupled to the processor. The input / output interface is configured to input and / or output information. The information includes at least one of instructions and data. The processor is configured to execute a computer program, such that the processing device performs the method performed by the first device or the second device in the above method embodiments.

[0308] An embodiment of the present application further provides a processing device including a processor and a memory, the memory configured to store a computer program, the processor configured to call the computer program from the memory and execute the computer program, such that the processing device performs the method performed by the first device or the second device in the above method embodiment.

[0309] It should be understood that the processing device can be one or more chips. For example, the processing device can be a field programmable gate array (FPGA), an application-specific integrated circuit (ASIC), a system on a chip (SoC), a central processing unit (CPU), a network processor (NP), a digital signal processor (DSP), a microcontroller unit (MCU), a programmable logic device (PLD), or another integrated chip.

[0310] In some implementation processes, the steps of the above method may be implemented by using hardware integrated logic circuits in a processor or by using instructions in the form of software. The steps of the methods disclosed with reference to the embodiments of the present application may be directly executed and completed by a hardware processor, or may be executed and completed by using a combination of hardware and software modules in a processor. The software modules may be located in a storage medium well-established in the art, such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, an electrically erasable programmable memory, or a register. The storage medium is located in the memory, and the processor reads information in the memory and completes the steps of the above method in cooperation with the hardware of the processor. To avoid repetition, details will not be described again here.

[0311] The processor in the embodiments of the present application may be an integrated circuit chip and have signal processing capabilities. In the implementation process, the steps in the above-described method embodiments can be implemented by using hardware integrated logic circuits in the processor or by using instructions in the form of software. The processor may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component. It may implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor may be a microprocessor, or the processor may be any conventional processor or similar. The steps in the methods disclosed with reference to the embodiments of the present application may be directly executed and completed by a hardware decode processor, or may be executed and completed by using a combination of hardware and software modules in the decode processor. The software modules may be located in a storage medium well-established in the art, such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, an electrically erasable programmable memory, or a register. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps in the above method in combination with the processor's hardware.

[0312] It will be understood that the memory in this embodiment of the present application may be volatile or non-volatile memory, or may include both volatile and non-volatile memory. Non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory may be random access memory (RAM) used as an external cache. By way of example and not limitation, many forms of RAM may be used, such as static random access memory (Static RAM, SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (Synchronous DRAM, SDRAM), double data rate synchronous dynamic random access memory (Double Data Rate SDRAM, DDR SDRAM), enhanced synchronous dynamic random access memory (Enhanced SDRAM, ESDRAM), synchronous link dynamic random access memory (Synchlink DRAM, SLDRAM), and direct Rambus dynamic random access memory (Direct Rambus RAM, DR RAM). It should be noted that memory in the systems and methods described herein includes, but is not limited to, these and any other suitable types of memory.

[0313] According to the method provided in the embodiment of the present application, the present application further provides a computer program product, which includes a computer program or a group of instructions, which, when executed on a computer, enables the computer to execute the method executed by the first device or the second device in the above method embodiment.

[0314] According to the method provided in the embodiment of the present application, the present application further provides a computer-readable storage medium, which stores a program, and when the program is executed on a computer, the computer can execute the method executed by the first device or the second device in the above embodiment of the method.

[0315] According to the method provided in the embodiment of the present application, the present application further provides a communication system, which may include the first device and the second device described above.

[0316] As used herein, terms such as "component," "module," and "system" are used to refer to computer-related entities, hardware, firmware, a combination of hardware and software, software, or software in execution. For example, a component may be, but is not limited to, a process running on a processor, a processor, an object, an executable, a thread of execution, a program, and / or a computer. As explained through the use of figures, both computing devices and applications running on a computing device may be components. One or more components may reside within a process and / or thread of execution, and components may be located on one computer and / or distributed between two or more computers. Furthermore, these components may execute from various computer-readable media that store various data structures. For example, components may communicate using local and / or remote processes and based on signals, e.g., comprising one or more data packets (e.g., data from two components interacting with another component in a local system, a distributed system, and / or across a network such as the Internet that interacts with other systems using signals).

[0317] Those skilled in the art may recognize that, in combination with the examples described in the embodiments disclosed herein, the units and algorithm steps may be implemented by electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software depends on the specific application and the design constraints of the technical solution. Those skilled in the art may use different methods to implement the described functions for each specific application, but the implementation should not be considered beyond the scope of this application.

[0318] For the purpose of convenience and concise description, for the detailed operation processes of the above systems, devices and units, it can be clearly understood by those skilled in the art to refer to the corresponding processes in the above method embodiments, and the details will not be described again here.

[0319] In the embodiments provided herein, the disclosed systems, devices, and methods may be implemented in other ways. For example, the described device embodiments are merely examples. For example, the division into multiple units is merely a logical division of function, and other divisions may be used in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not implemented. In addition, the shown or described interconnections or direct connections or communication connections may be implemented using some interfaces. Indirect connections or communication connections between devices or units may be implemented in electronic, mechanical, or other forms.

[0320] Units described as separate parts may or may not be physically separate, and parts shown as units may or may not be physical units, located in one place or distributed over multiple network units. Some or all of the units may be selected based on actual requirements to achieve the objectives of the solutions of the embodiments.

[0321] In addition, the functional units in the embodiments of the present application may be integrated into one processing unit, each of the units may exist physically alone, or two or more units may be integrated into one unit.

[0322] When a function is implemented in the form of a software functional unit and sold or used as an independent product, the function may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application or a part that essentially contributes to a part of the technical solution may be embodied in the form of a software product. A computer software product is stored in a storage medium and includes a plurality of instructions for enabling a computer device (which may be a personal computer, a server, a network device, or the like) to execute all or some of the steps of the method in the embodiments of the present application. The above-mentioned storage medium includes any medium that can store program code, such as a USB flash drive, a removable hard disk, a ROM, a RAM, a magnetic disk, or an optical disk.

[0323] The above description is merely a specific implementation of the present application and is not intended to limit the scope of protection of the present application. Any modifications or replacements that can be easily thought of by those skilled in the art within the technical scope disclosed in the present application shall be included in the scope of protection of the present application. Therefore, the scope of protection of the present application shall be subject to the scope of protection of the claims.

Claims

1. 1. A data compression and transmission method comprising: Obtaining M first data, where one sub-data in the first data corresponds to one first sparse matrix, the first sparse matrix represents one corresponding sub-data in the first data based on a first dictionary matrix, and the first dictionary matrix includes characteristics of the M sub-data corresponding to the M first data respectively; and outputting compressed data of the M first sparse matrices, where M is an integer greater than 1; A method for providing

2. For the pth first data and the qth first data among the M first data, the similarity between one sub-data in the pth first data and one sub-data in the qth first data is greater than or equal to a first similarity threshold; and Both p and q are positive integers, and p is not equal to q; The method of claim 1.

3. The M first data are data for M time units, respectively, and each first data includes N sub-data obtained through division based on a spatial positional relationship, where N is a positive integer; or The M first data are data in h time units, and the data in the h time units are sorted based on a spatial positional relationship to obtain the M first data, where one first data includes N sub-data, and h is a positive integer; and Among the M first data, a similarity between a spatial position of one subdata in the pth first data and a spatial position of one subdata in the qth first data is greater than or equal to a first similarity threshold; 3. The method according to claim 1 or 2.

4. The method further comprises: The k-th first data Y among the M first data ｋ Regarding the kth first data Y ｋ decomposing one sub-data in into the first dictionary matrix and k first sparse matrices, where k is a positive integer less than or equal to M; and determining M-1 first sparse matrices other than the k first sparse matrix based on the first dictionary matrix; The method of any one of claims 1 to 3, comprising:

5. The step of outputting the compressed data of the M first sparse matrices comprises: determining a first matrix based on the M first sparse matrices; performing a low-rank approximation on the first matrix to obtain first compressed data of the M first sparse matrices; and outputting the first compressed data 5. The method of claim 1, comprising:

6. The step of determining the first matrix based on the M first sparse matrices comprises: combining the M first sparse matrices to obtain the first matrix; or performing data compression on at least one of the M first sparse matrices; and combining the M first sparse matrices obtained through the data compression to obtain the first matrix. The method of claim 5 , comprising:

7. The step of performing data compression on the at least one of the M first sparse matrices comprises: for one of the M first sparse matrices, setting a value of a first element in the one first sparse matrix to a first value based on first position indication information; where: the first position indication information indicates a position of an element in the first sparse matrix that can represent one corresponding sub-data, and the first element cannot represent one sub-data in the first data; The method of claim 6.

8. The step of performing low-rank approximation on the first matrix to obtain the first compressed data of the M first sparse matrices comprises: performing singular value decomposition on the first matrix to obtain K feature values ​​and feature vectors corresponding to the K feature values, respectively, wherein the K feature values ​​and the feature vectors corresponding to the K feature values ​​respectively represent the first matrix; and K of the K characteristic values ０ characteristic values, and the K ０ using the characteristic vectors corresponding to the M characteristic values ​​as the first compressed data of the M first sparse matrices.

8. The method of any one of claims 5 to 7, comprising:

9. outputting second compressed data of the M first sparse matrices, wherein the second compressed data is the K characteristic values ​​of the K characteristic values; ０ K other than the characteristic values ​​of １ characteristic values, and the K １ The method of claim 8 , further comprising: including a feature vector corresponding to each of the feature values.

10. The step of outputting the compressed data of the M first sparse matrices comprises: outputting first residual information, wherein the first residual information is determined based on information about a j-th first sparse matrix and information about an i-th first sparse matrix among the M first sparse matrices; where: i is less than j, and both i and j are positive integers; 5. The method according to any one of claims 1 to 4.

11. The step of outputting the first residual information comprises: determining a similarity between the jth first sparse matrix and the ith first sparse matrix in the M first sparse matrices; and outputting the first residual information when the similarity is less than or equal to a second similarity threshold. The method of claim 10, comprising:

12. The method of claim 10 or 11, wherein the information about the i-th first sparse matrix is ​​obtained by decompressing compressed data of the i-th first sparse matrix.

13. 13. The method of claim 10, wherein the first residual information comprises a first sequence of residual elements, the first sequence of residual elements representing a residual matrix between the information for the jth first sparse matrix and the information for the ith first sparse matrix, where i is less than j and both i and j are positive integers.

14. 14. The method of claim 13, wherein the first residual information further comprises second position indication information, the second position indication information indicating that an absolute value is greater than or equal to in the residual matrix between the information for the jth first sparse matrix and the information for the ith first sparse matrix.

15. outputting second residual information, wherein the second residual information is determined based on the information for the jth first sparse matrix and the information for the ith first sparse matrix, the second residual information includes third position indication information and a second sequence of residual elements, the third position indication information indicating residual elements in the residual matrix between the information for the jth first sparse matrix and the information for the ith first sparse matrix that are less than or equal to the first residual threshold and greater than or equal to a second residual threshold, and the second sequence of residual elements includes residual elements in the residual matrix that are less than or equal to the first residual threshold and greater than or equal to the second residual threshold. The method of any one of claims 10 to 14, further comprising:

16. The method further comprises: transmitting the first dictionary matrix; or receiving the first dictionary matrix 16. The method of any one of claims 1 to 15, comprising:

17. The step of outputting the first compressed data comprises: performing a compression process on the first compressed data, wherein the compression process includes quantization and / or entropy coding; and outputting the first compressed data obtained through the compression process; 10. The method of any one of claims 5 to 9, comprising:

18. The step of outputting the first residual information comprises: performing a compression process on the first residual information, wherein the compression process includes quantization and / or entropy coding; and Outputting the first residual information obtained through the compression process. The method of claim 11 , comprising:

19. transmitting a first indication; or receiving first indication information wherein The first indication information: The number of properties of the low-rank approximation, K ０ ; Number of first data M; a capacity threshold, wherein the capacity threshold is used to determine whether an element in the first sparse matrix can represent one corresponding sub-data in the first data; a first residual threshold, where the first residual threshold is used to determine a first residual element sequence, where the first residual element sequence represents a residual matrix between the information for the jth first sparse matrix and the information for the ith first sparse matrix, where i is less than j and both i and j are positive integers; a proportion of elements in one of the M first sparse matrices that can represent one corresponding subdata in the first data; data loss of the first compressed data for the M first sparse matrices, wherein the first compressed data is determined based on the M first sparse matrices; compression process parameters, where the compression process includes quantization and / or entropy coding, and the compression process parameters include at least one of a quantization precision, a quantization codebook, and a coding scheme; and whether to transmit first residual information, wherein the first residual information is determined based on the information about the jth first sparse matrix and the information about the ith first sparse matrix among the M first sparse matrices; 19. The method of claim 1, wherein the method indicates at least one of:

20. Based on the first time-frequency resources of the M first data: the capacity threshold, wherein the capacity threshold is used to determine whether the element in the first sparse matrix can represent one corresponding sub-data in the first data; a first residual threshold, where the first residual threshold is used to determine the first residual element sequence, the first residual element sequence representing the residual matrix between the information for the jth first sparse matrix and the information for the ith first sparse matrix, where i is less than j and both i and j are positive integers; and the compression processing parameters, wherein the compression processing includes quantization and / or entropy coding, and the compression processing parameters include at least one of the quantization precision, the quantization codebook, and the coding method; determining at least one of 20. The method of any one of claims 1 to 19, further comprising:

21. 21. The method of claim 20, further comprising receiving first configuration information, wherein the first configuration information is used to configure the first time-frequency resource.

22. 22. The method of claim 21, further comprising: sending a compress and transmit request, wherein the compress and transmit request carries data types of the M first data, the data types including point cloud data and / or artificial intelligence AI data.

23. transmitting a second indication; or receiving second indication information, The second indication information: The number of properties of the low-rank approximation, K １ and a second residual threshold, wherein the second residual threshold is used in combination with the first residual threshold to determine a second residual element sequence, the second residual element sequence representing the residual matrix between the information for the jth first sparse matrix and the information for the ith first sparse matrix, where i is less than j and both i and j are positive integers; Indicate at least one of 23. The method of any one of claims 1 to 22, further comprising:

24. determining the second residual threshold based on second time-frequency resources of the M first data, wherein the second residual threshold is used in combination with the first residual threshold to determine the second residual element sequence, the second residual element sequence representing the residual matrix between the information about the jth first sparse matrix and the information about the ith first sparse matrix, where i is less than j and both i and j are positive integers; 24. The method of any one of claims 1 to 23, further comprising:

25. receiving second configuration information, wherein the second configuration information is used to configure the second time-frequency resource.

25. The method of claim 24, further comprising:

26. 1. A data compression and transmission method comprising: receiving compressed data of M first sparse matrices, where the first sparse matrices represent corresponding sub-data in the first data based on a first dictionary matrix, the first dictionary matrix includes characteristics of the M sub-data corresponding to the M first data respectively, and one sub-data in the first data corresponds to one first sparse matrix; and outputting decompression information based on the compressed data A method for providing

27. For the pth first data and the qth first data among the M first data, the similarity between one sub-data in the pth first data and one sub-data in the qth first data is greater than or equal to a first similarity threshold; and Both p and q are positive integers, and p is not equal to q; 27. The method of claim 26.

28. The M first data are data for M time units, respectively, and each first data includes N sub-data obtained through division based on a spatial positional relationship, where N is a positive integer; or The M first data are data in h time units, and the data in the h time units are sorted based on a spatial positional relationship to obtain the M first data, where one first data includes N sub-data, and h is a positive integer; and a similarity between a spatial position of one subdata in the pth first data and a spatial position of one subdata in the qth first data among the M first data is greater than or equal to the first similarity threshold; 28. The method of claim 26 or 27.

29. The compressed data includes first compressed data, and the step of outputting the decompression information based on the compressed data comprises: performing low-rank matrix reconstruction based on the first compressed data to obtain a first matrix; determining the M first sparse matrices based on the first matrix; constructing the M first data based on the M first sparse matrices and the first dictionary matrix; and Outputting the M first data 29. The method of any one of claims 26 to 28, comprising:

30. The step of determining the M first sparse matrices based on the first matrix comprises: dividing the first matrix to obtain the M first sparse matrices; or dividing the first matrix to obtain decompression information for the M first sparse matrices, and performing data decompression on at least one of the M first sparse matrices based on the information for the M first sparse matrices.

30. The method of claim 29, comprising:

31. The step of performing data decompression on the at least one of the M first sparse matrices comprises: performing data decompression on one of the M first sparse matrices based on first position indication information, the first position indication information indicates a position of an element in the first sparse matrix that can represent one corresponding subdata; 31. The method of claim 30, comprising:

32. The first compressed data is K ０ characteristic values, and the K ０ the step of performing low-rank matrix reconstruction based on the first compressed data to obtain the first matrix includes: The above K ０ characteristic values, and the K ０ performing low-rank matrix reconstruction based on the feature vectors corresponding to the feature values ​​to obtain the first matrix; 32. The method of any one of claims 29 to 31, comprising:

33. The compressed data further includes second compressed data, １ characteristic values, and the K １ the characteristic vectors corresponding to the K characteristic values, １ The characteristic values ​​are ０ 33. The method of claim 32, wherein the characteristic value is different from the characteristic value.

34. The compressed data includes first residual information, and the first residual information is determined based on information about a j-th first sparse matrix and information about an i-th first sparse matrix among the M first sparse matrices, and the step of outputting the decompression information based on the compressed data comprises: constructing the jth first sparse matrix based on the first residual information and the ith first sparse matrix; constructing the M first data based on the jth first sparse matrix; and outputting the M first data, i is less than j, and both i and j are positive integers; Including, 29. The method of any one of claims 26 to 28.

35. 35. The method of claim 34, wherein the information about the i-th first sparse matrix is ​​obtained by decompressing compressed data of the i-th first sparse matrix.

36. 36. The method of claim 34 or 35, wherein the first residual information includes a first residual element sequence, the first residual element sequence representing a residual matrix between the information for the jth first sparse matrix and the information for the ith first sparse matrix, where i is less than j and both i and j are positive integers.

37. 37. The method of claim 36, wherein the first residual information further comprises second position indication information, the second position indication information indicating a position of a residual element in the residual matrix between the information for the jth first sparse matrix and the information for the ith first sparse matrix, the absolute value of which is greater than or equal to a first residual threshold.

38. the compressed data further includes second residual information, the second residual information being determined based on the information about the jth first sparse matrix and the information about the ith first sparse matrix; the second residual information includes third position indication information and a second sequence of residual elements, the third position indication information indicating residual elements in the residual matrix between the information for the jth first sparse matrix and the information for the ith first sparse matrix that are less than or equal to the first residual threshold and greater than or equal to a second residual threshold, and the second sequence of residual elements includes residual elements in the residual matrix that are less than or equal to the first residual threshold and greater than or equal to the second residual threshold.

38. The method of any one of claims 34 to 37.

39. The method further comprises: receiving the first dictionary matrix; or Transmitting the first dictionary matrix 39. The method of any one of claims 26 to 38, comprising:

40. receiving a first indication; or transmitting first instruction information wherein The first indication information: The number of properties of the low-rank approximation, K ０ ; Number of first data M; a capacity threshold, wherein the capacity threshold is used to determine whether an element in the first sparse matrix can represent one corresponding sub-data in the first data; a first residual threshold, where the first residual threshold is used to determine a first residual element sequence, where the first residual element sequence represents a residual matrix between information about the jth first sparse matrix and the information about the ith first sparse matrix, where i is less than j and both i and j are positive integers; a proportion of elements in one of the M first sparse matrices that can represent one corresponding subdata in the first data; data loss of the first compressed data for the M first sparse matrices, wherein the first compressed data is determined based on the M first sparse matrices; compression process parameters, where the compression process includes quantization and / or entropy coding, and the compression process parameters include at least one of a quantization precision, a quantization codebook, and a coding scheme; and whether to transmit first residual information, wherein the first residual information is determined based on the information about the jth first sparse matrix and the information about the ith first sparse matrix among the M first sparse matrices; 40. The method of any one of claims 26 to 39, wherein the method indicates at least one of:

41. transmitting first configuration information, wherein the first configuration information is used to configure first time-frequency resources of the M first data; 41. The method of any one of claims 26 to 40, further comprising:

42. 42. The method of claim 41, further comprising receiving a compression and transmission request, wherein the compression and transmission request carries data types of the M first data, the data types including point cloud data and / or artificial intelligence AI data.

43. receiving the transmitted second instruction information; or transmitting second indication information, The second indication information: The number of properties of the low-rank approximation, K １ and the second residual threshold, wherein the second residual threshold is used in combination with the first residual threshold to determine the second residual element sequence, the second residual element sequence representing the residual matrix between the information for the jth first sparse matrix and the information for the ith first sparse matrix, where i is less than j and both i and j are positive integers; Indicate at least one of 43. The method of any one of claims 26 to 42, further comprising:

44. 44. The method of any one of claims 26 to 43, further comprising receiving second configuration information, wherein the second configuration information is used to configure second time-frequency resources of the M first data.

45. A communications device comprising a module configured to perform the method of any one of claims 1 to 25 or comprising a module configured to perform the method of any one of claims 26 to 44.

46. 45. A communications device comprising a processor, the processor being configured to perform the method of any one of claims 1 to 44 by executing a computer program or by using logic circuitry.

47. 45. A computer readable storage medium configured to store computer program instructions, the computer program causing a computer to perform the method of any one of claims 1 to 44.

48. A computer program product comprising computer program instructions, the computer program instructions causing a computer to carry out a method according to any one of claims 1 to 44.

Images