A P-LDPC code encoding method, device, medium and program product

Abstract

The present invention provides a coding method, device, medium and program product for a P-LDPC code. The method includes: S1 converting the original film image matrices of multiple P-LDPC codes into test units; S2 defining a mean shift function; S3 calculating the maximum value of the mean shift function to obtain a first initial centroid; S4 calculating a first distortion value of the first initial centroid; S5 calculating a second distortion value of a second test unit closest to the first initial centroid; S6 comparing the first distortion value with the second distortion value; if the second distortion value is less than or equal to the first distortion value, the second test unit is used as the first initial centroid, and it is determined whether the second distortion value is less than a threshold value. If so, the original film image matrix reconstructed by the first initial centroid is output; if not, the process goes to step S3; if the second distortion value is greater than the first distortion value, the test unit closest to the first initial centroid is selected from other test units as the second test unit, and the process goes to step S5. The above scheme reduces the coding design complexity of the original film image low-density parity check code.

Description

Technical Field

[0001] The present invention relates to the field of information and communication technology, and in particular to a P-LDPC code encoding method, device, medium and program product. Background Art

[0002] In the field of signal processing, lossy data compression can effectively save bandwidth resources for signal source storage and transmission, making it a key technology for achieving green communications. In the Internet of Things (IoT) scenario, wireless sensor networks are widely deployed to collect and interact with environmental information. These sensor nodes are often located in complex environments and require long-term operation, making their data collection efficiency and transmission performance critical communication metrics. Lossy data compression provides a simple and direct signal processing method that simplifies data and reduces transmission resources, effectively improving collection and transmission efficiency. This method is particularly suitable for processing large data sets such as voice, images, and video. Lossy compression converts the original analog signal into a more compact digital form while maintaining sufficient information quality. Lossy source coding achieves compression by removing redundant portions of the signal, ensuring the validity and usability of the compressed data. Furthermore, by compressing the data as a preprocessing step, the complexity of subsequent communication processing and the required hardware costs can be significantly reduced.

[0003] Among the technologies required for data compression, the existing technology proposes a lossy source coding scheme based on the Protograph Low-density Parity-check (P-LDPC) code, which realizes efficient source modulus expression. At the same time, relevant research shows that the rate-distortion performance of lossy source coding can be further improved by optimizing the coding of the Protograph Matrix of the P-LDPC code. However, the optimized target P-LDPC code is obtained by searching using the Differential Evolution (DE) algorithm. The DE algorithm is a global brute force search method, and its search complexity grows exponentially. Using the DE algorithm to optimize the Protograph Matrix of the P-LDPC code will lead to low efficiency in the optimization design of the communication encoder, consuming a lot of manpower, computing power and time.

[0004] Therefore, there is currently a lack of high-efficiency coding technology based on optimized design of P-LDPC codes, resulting in slow updates and improvements to encoders in practical applications. This makes it difficult to quickly iterate encoders for processing data from different sources, increasing the power consumption and cost of information processing and communication. In practical applications, higher hardware resource development and human resource maintenance are required. Summary of the Invention

[0005] The embodiments of the present invention provide a P-LDPC code encoding method, apparatus, medium, and program product to reduce the encoding design complexity of the original low-density parity-check code.

[0006] In order to achieve the above object, on the one hand, a P-LDPC code encoding method is provided, the method comprising the following steps:

[0007] S1, converting the first original film image matrices of the plurality of P-LDPC codes into a plurality of column vectors, and preparing a test data set by using the plurality of column vectors as test units, wherein all the test units in the test data set have the same dimension;

[0008] S2, defining the mean shift function and the initial centroid of the test data set;

[0009] The mean shift function is: Among them, N represents the density of the test unit, represents the nuclear radius, represents the kernel function, N represents the number of test units within the kernel radius R, c0 represents the initial center of mass, c i represents a variable centroid, which is a test unit within the core radius R;

[0010] S3, calculating a first mean shift direction of the variable centroid within a preset first core radius, finding a maximum value of the mean shift function based on the first mean shift direction through gradient calculation, and updating the initial centroid based on the first test unit to which the maximum value belongs, to obtain a first initial centroid;

[0011] The gradient is calculated as: in, Represents the kernel function The derivative of , at this time, the value of R is the preset first nuclear radius;

[0012] S4, reconstructing the first initial centroid into a second original membrane image matrix, and calculating a distortion value of the second original membrane image matrix according to a mutual information iterative convergence algorithm to obtain a first distortion value;

[0013] S5, calculating a second mean shift direction of the variable centroid within a preset second core radius, and selecting a test unit closest to the first initial centroid in the second mean shift direction as a second test unit;

[0014] S6, reconstructing the second test unit into the third original film image matrix, and calculating the distortion value of the third original film image matrix according to the mutual information iterative convergence algorithm to obtain a second distortion value; wherein the first core radius is greater than the second core radius;

[0015] S7, compare the first distortion value and the second distortion value; if the second distortion value is less than or equal to the first distortion value, update the first initial centroid to the second test unit, and further determine whether the second distortion value is less than a preset first threshold value. If so, output the original membrane image matrix reconstructed by the current first initial centroid; if not, go to step S4; if the second distortion value is greater than the first distortion value, then for other test units within the second core radius except the current second test unit, select the test unit closest to the first initial centroid as the second test unit again, and go to step S6.

[0016] Preferably, in the encoding method of the P-LDPC code, in step S3, the maximum value of the mean shift function is found by gradient calculation:

[0017] Calculate the first gradient value of the first test unit that is closest to the initial centroid in the first mean shift direction, and determine whether the first gradient value is greater than a preset gradient threshold. If so, update the first test unit to the initial centroid, and calculate and determine the first gradient value again; if not, determine the value of the mean shift function calculated using the first test unit as the maximum value.

[0018] Preferably, in the encoding method of the P-LDPC code, in step S3, the first mean shift direction and the second mean shift direction are determined by the following formula:

[0019]

[0020] Among them, E(c i ) represents the mean shift direction of the variable centroid, and g() represents the negative of the derivative of the kernel function.

[0021] Preferably, in the encoding method of the P-LDPC code, in step S3, the test unit is represented as:

[0022]

[0023] Where c represents the test unit, m represents the number of rows, n represents the number of columns, b represents an integer, and T represents transpose;

[0024] The original membrane map matrix reconstructed by the test unit is expressed as:

[0025]

[0026] Preferably, in the encoding method of the P-LDPC code, in step S3, the mutual information iterative convergence algorithm is:

[0027]

[0028] in, represents the mutual information iterative convergence algorithm, d0 is the distortion value, R B represents the code rate of the original film image matrix B to which the distortion value d0 belongs, h(x) is the entropy of the variable x, and e represents the natural base.

[0029] Preferably, in the encoding method of the P-LDPC code, in the step S3, when there is no test unit with a distortion value less than the first threshold within the second core radius, the process proceeds to step S3.

[0030] On the other hand, a P-LDPC code encoding device includes a memory and a processor, wherein the memory stores at least one program, and the at least one program is executed by the processor to implement any of the P-LDPC code encoding methods described above.

[0031] On the other hand, a computer-readable storage medium is provided, wherein the storage medium stores at least one program, and the at least one program is executed by a processor to implement any of the P-LDPC code encoding methods described above.

[0032] In yet another aspect, a computer program product includes a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-described P-LDPC code encoding methods.

[0033] The above technical solution has the following technical effects:

[0034] The technical solution of the embodiment of the present invention decomposes the original film image matrix of the original film image low-density parity check code (P-LDPC code) into multiple column vectors as test units, sets the initial centroid and mean shift function, and then gradually updates the centroid to near the maximum point of the mean shift function by calculating the mean shift direction of the centroid under the first core radius, thereby obtaining the optimized centroid position. Subsequently, the optimized centroid is reconstructed into a new original film image matrix, and its distortion value is calculated using the mutual information iterative convergence algorithm. The distortion value of the test unit within the second core radius is compared to determine whether to update the centroid position, and the iteration is continued until the distortion value is lower than the preset threshold or cannot be further reduced. The above technical solution can automatically find the optimal original film image matrix, significantly improve the performance indicators of the P-LDPC code, reduce the bit error rate of the P-LDPC code, and improve the decoding efficiency of the P-LDPC code.

[0035] In a further embodiment, by setting the first kernel radius and the second kernel radius in combination with the mean shift function to realize iterative update of the center of mass, the design of the low-density parity-check code of the original membrane image is optimized, so that the encoder for processing different source data can be updated and improved faster, which helps to achieve rapid adaptation of the encoder to different application scenarios; by calculating the minimum distortion of the reconstructed original membrane image matrix through the mutual information iterative convergence algorithm, the original membrane image matrix can be found faster, thereby improving the efficiency of the coding design. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 1 is a flow chart of a P-LDPC code encoding method according to an embodiment of the present invention;

[0037] Figure 2 1 is a schematic diagram of a process for implementing an MDSC algorithm in a P-LDPC code encoding method according to an embodiment of the present invention;

[0038] Figure 3 FIG. 4 is a diagram of a P-LDPC code encoding device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0039] To further illustrate various embodiments, the present invention is provided with accompanying drawings. These drawings form part of the present disclosure and are primarily used to illustrate the embodiments and, in conjunction with the relevant description in the specification, to explain the operating principles of the embodiments. By referring to these drawings, one of ordinary skill in the art will understand other possible embodiments and the advantages of the present invention. The components in the figures are not drawn to scale, and similar reference numerals are generally used to represent similar components.

[0040] The present invention will now be further described with reference to the accompanying drawings and specific embodiments.

[0041] Example 1:

[0042] In order to reduce the bit error rate of the P-LDPC code and further improve the decoding efficiency of the P-LDPC code, this embodiment provides a P-LDPC code encoding method. Figure 1 FIG. 1 is a flow chart of a P-LDPC code encoding method according to an embodiment of the present invention. Figure 1 As shown, the method includes the following steps:

[0043] S1, converting the first original film image matrices of the plurality of P-LDPC codes into a plurality of column vectors, and preparing a test data set by using the plurality of column vectors as test units, wherein all the test units in the test data set have the same dimension;

[0044] S2, defines the mean shift function and the initial centroid of the test dataset;

[0045] The mean shift function is: Among them, N represents the density of the test unit, represents the nuclear radius, Represents the kernel function, N represents the number of test units within the kernel radius R, c0 represents the initial center of mass, c i represents the variable centroid, which is the test unit within the core radius R;

[0046] S3, calculating a first mean shift direction of the variable centroid within a preset first core radius, finding a maximum value of the mean shift function based on the first mean shift direction through gradient calculation, and updating the initial centroid based on the first test unit to which the maximum value belongs, to obtain a first initial centroid;

[0047] The gradient is calculated as: in, Represents the kernel function The derivative of The value of is the preset first core radius;

[0048] Preferably, the first gradient value of the first test unit that is closest to the initial centroid in the first mean shift direction is calculated, and it is determined whether the first gradient value is greater than a preset gradient threshold. If so, the first test unit is updated to the initial centroid, and the first gradient value is calculated and determined again; if not, the value of the mean shift function calculated using the first test unit is determined to be the maximum value.

[0049] S4, reconstructing the first initial centroid into a second original membrane image matrix, calculating the distortion value of the second original membrane image matrix according to a mutual information iterative convergence algorithm, and obtaining a first distortion value;

[0050] S5, calculating a second mean shift direction of the variable centroid within a preset second core radius, and selecting a test unit closest to the first initial centroid in the second mean shift direction as a second test unit;

[0051] S6, reconstructing the second test unit into a third original membrane image matrix, and calculating the distortion value of the third original membrane image matrix according to the mutual information iterative convergence algorithm to obtain a second distortion value; wherein the first kernel radius is greater than the second kernel radius;

[0052] S7, compare the first distortion value and the second distortion value; if the second distortion value is less than or equal to the first distortion value, update the first initial centroid to the second test unit, and further determine whether the second distortion value is less than a preset first threshold value. If so, output the original membrane image matrix reconstructed by the current first initial centroid; if not, go to step S4; if the second distortion value is greater than the first distortion value, then for other test units within the second core radius except the current second test unit, select the test unit closest to the first initial centroid as the second test unit again, and go to step S6.

[0053] Preferably, when there is no test unit with a distortion value less than the first threshold within the second core radius, the process proceeds to step S3.

[0054] Preferably, the first mean shift direction and the second mean shift direction are determined by the following formula:

[0055]

[0056] Among them, E(c i ) represents the mean shift direction of the variable centroid, and g() represents the negative of the derivative of the kernel function.

[0057] Example 2:

[0058] In order to realize high-efficiency coding technology based on the optimized design of P-LDPC code and reduce the optimization design complexity of the original membrane matrix of the P-LDPC code, this embodiment proposes a minimum distortion shift clustering (MDSC) method to find the target original membrane matrix of the source P-LDPC code. The basic principle of the MDSC method is single-target clustering, that is, clustering the test data set of the test unit containing multiple P-LDPC codes into a target P-LDPC code with minimal distortion. The MDSC method can effectively reduce the power consumption and labor cost of the communication system design of information compression, and further realize the optimization of the communication system to build a more green and environmentally friendly wireless sensor network. The method includes the following contents:

[0059] The original film graph matrix of the P-LDPC code is converted into a column vector as a test unit to prepare a test data set, and in one test data set, the dimensions of all test units are the same.

[0060] Define the initial centroid and mean shift function in, represents the cell density, is the nuclear radius, r1 is the larger nuclear radius set in advance, r2 is the smaller nuclear radius set in advance, r1 and r2 both belong to Nuclear radius, is the kernel function, N represents the number of test units within the kernel radius R, c0 is the initial centroid, c i is the variable center of mass.

[0061] Complete the maximum density clustering. In order to obtain the maximum value of the function f(c0), the gradient of the mean shift function is calculated, and the initial centroid is updated by the variable centroid, where:

[0062]

[0063] Represents the kernel function The derivative of , the nuclear radius R is r1.

[0064] In a specific implementation, the first gradient value of the first test unit closest to the initial centroid is calculated, and it is determined whether the first gradient value is greater than a preset gradient threshold. If so, the first test unit is updated to the initial centroid, and the first gradient value is calculated and determined again; if not, the value of the mean shift function calculated using the first test unit is determined to be the maximum value.

[0065] Complete the minimum distortion clustering. When the kernel radius is r2, the initial centroid is updated within the kernel radius of r1 through the mean shift function. c represents the test unit, m represents the number of rows, n represents the number of columns, b represents an integer, and T represents transposition. The updated initial centroid is reconstructed into the original membrane map matrix The minimum distortion original membrane matrix is ​​calculated by the Mutual Information Iteration Convergence (MIIC) algorithm. Specifically, It is the objective function of the original membrane image matrix B with minimum distortion d0, expressed as Calculated as Get the minimum distortion, where R B represents the code rate of the original membrane image matrix B, and h(x) is the entropy of the unknown variable x.

[0066] In a specific implementation, the updated initial centroid is reconstructed into a second original membrane map matrix, and the distortion value of the second original membrane map matrix is ​​calculated according to the mutual information iterative convergence algorithm to obtain a first distortion value; the test unit closest to the updated initial centroid is selected as the second test unit, the second test unit is reconstructed into a third original membrane map matrix, and the distortion value of the third original membrane map matrix is ​​calculated according to the mutual information iterative convergence algorithm to obtain a second distortion value; the first distortion value is compared with the second distortion value, if the second distortion value is less than or equal to the first distortion value, the initial centroid is updated to the second test unit, and it is further determined whether the second distortion value is less than a preset first threshold. If so, the original membrane map matrix reconstructed by the current initial centroid is output; if the second distortion value is greater than the first distortion value, then for other test units within the core radius r2 except the current second test unit, the test unit closest to the initial centroid is selected as the second test unit again, and the judgment is recalculated and compared.

[0067] The accuracy of the MDSC method is tested using minimum distortion. Based on the clustering characteristics of test units with the same characteristics in the test dataset, the compression rate increases as the dimension of the test unit increases, and the distortion of the experimental results decreases with the increase of the dimension, thereby completing the experimental verification of the output accuracy of the MDSC method.

[0068] Preferably, the maximum density clustering and the minimum distortion clustering update the initial centroid according to the calculated mean shift direction, where:

[0069]

[0070] E(c i ) represents the mean shift direction of the variable centroid, and g() represents the negative of the derivative of the kernel function.

[0071] The Minimum Distortion Shift Clustering (MDSC) method proposed in this embodiment designs a two-step clustering based on the clustering characteristics of the data set, effectively solving the problem of high coding design complexity and can be further applied to channel coding optimization based on the principle of coding "duality". Figure 2 FIG. 1 is a schematic diagram of the process of implementing the MDSC algorithm in the encoding method of the P-LDPC code according to an embodiment of the present invention. Figure 2 As shown, the initial centroid moves according to the calculated mean shift direction and a vector, and the centroid with specific minimum distortion in the test data set is found through maximum density clustering and minimum distortion clustering.

[0072] Example 3:

[0073] The present invention also provides a P-LDPC code encoding device, such as Figure 3As shown, the device includes a processor 301, a memory 302, a bus 303, and a computer program stored in the memory 302 and executable on the processor 301. The processor 301 includes one or more processing cores. The memory 302 is connected to the processor 301 via the bus 303. The memory 302 is used to store program instructions. When the processor executes the computer program, the steps in the above-mentioned method embodiment of the first embodiment of the present invention are implemented.

[0074] Furthermore, as an executable solution, the P-LDPC code encoding device can be a computer unit, which can be a computing device such as a desktop computer, laptop, PDA, or cloud server. The computer unit may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above-described computer unit structure is merely an example and does not constitute a limitation of the computer unit. The computer unit may include more or fewer components than described above, or a combination of certain components, or different components. For example, the computer unit may also include input and output devices, network access devices, buses, etc., but this is not limited in the embodiments of the present invention.

[0075] Furthermore, as an executable solution, the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the computer unit and connects various parts of the entire computer unit using various interfaces and lines.

[0076] The memory can be used to store the computer programs and / or modules, and the processor implements the various functions of the computer unit by running or executing the computer programs and / or modules stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system and at least one application required for a function; the data storage area can store data created based on the use of the mobile phone, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0077] Example 4:

[0078] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method in the embodiment of the present invention are implemented.

[0079] If the module / unit integrated in the computer unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the process in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, it can implement the steps of the above-mentioned various method embodiments. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM) and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction.

[0080] Embodiment 5:

[0081] The present invention also provides a computer program product, comprising a computer program, which implements the steps of the method described above when executed by a processor.

[0082] Although the present invention has been particularly shown and described in conjunction with preferred embodiments, it will be understood by those skilled in the art that various changes in form and details may be made to the present invention without departing from the spirit and scope of the invention as defined in the appended claims, and all such changes are within the scope of protection of the present invention.

Claims

1. A P-LDPC code encoding method, characterized in that: The following steps are involved: S1, converting the first original film image matrices of the plurality of P-LDPC codes into a plurality of column vectors, and preparing a test data set by using the plurality of column vectors as test units, wherein all the test units in the test data set have the same dimension; S2, defining the mean shift function and the initial centroid of the test data set; The mean shift function is: Among them, N represents the density of the test unit, represents the nuclear radius, represents the kernel function, N represents the number of test units within the kernel radius R, c0 represents the initial center of mass, c i represents a variable centroid, which is a test unit within the core radius R; S3, calculating a first mean shift direction of the variable centroid within a preset first core radius, finding a maximum value of the mean shift function based on the first mean shift direction through gradient calculation, and updating the initial centroid based on the first test unit to which the maximum value belongs, to obtain a first initial centroid; The gradient is calculated as: in, Represents the kernel function The derivative of The value of is the preset first core radius; S4, reconstructing the first initial centroid into a second original membrane image matrix, and calculating a distortion value of the second original membrane image matrix according to a mutual information iterative convergence algorithm to obtain a first distortion value; S5, calculating a second mean shift direction of the variable centroid within a preset second core radius, and selecting a test unit closest to the first initial centroid in the second mean shift direction as a second test unit; S6, reconstructing the second test unit into a third original membrane image matrix, and calculating the distortion value of the third original membrane image matrix according to the mutual information iterative convergence algorithm to obtain a second distortion value; wherein the first core radius is greater than the second core radius; S7, compare the first distortion value and the second distortion value; if the second distortion value is less than or equal to the first distortion value, update the first initial centroid to the second test unit, and further determine whether the second distortion value is less than a preset first threshold value. If so, output the original membrane image matrix reconstructed by the current first initial centroid; if not, go to step S4; if the second distortion value is greater than the first distortion value, then for other test units within the second core radius except the current second test unit, select the test unit closest to the first initial centroid as the second test unit again, and go to step S6.

2. The encoding method of P-LDPC code according to claim 1, characterized in that In step S3, the maximum value of the mean shift function is found by gradient calculation: Calculate the first gradient value of the first test unit that is closest to the initial centroid in the first mean shift direction, and determine whether the first gradient value is greater than a preset gradient threshold. If so, update the first test unit to the initial centroid, and calculate and determine the first gradient value again; if not, determine the value of the mean shift function calculated using the first test unit as the maximum value.

3. The encoding method of P-LDPC code according to claim 1, characterized in that The first mean shift direction and the second mean shift direction are determined by the following formula: Among them, E(c i ) represents the mean shift direction of the variable centroid, and g() represents the negative of the derivative of the kernel function.

4. The encoding method of P-LDPC code according to claim 1, characterized in that Represent a test unit as: Where c represents the test unit, m represents the number of rows, n represents the number of columns, b represents an integer, and T represents transpose; The original membrane map matrix reconstructed by the test unit is expressed as:

5. The encoding method of P-LDPC code according to claim 1, characterized in that The mutual information iterative convergence algorithm is: in, represents the mutual information iterative convergence algorithm, d0 is the distortion value, R B represents the code rate of the original film image matrix B to which the distortion value d0 belongs, h(x) is the entropy of the variable x, and e represents the natural base.

6. The encoding method of P-LDPC code according to claim 1, characterized in that When there is no test unit with a distortion value smaller than the first threshold within the second core radius, the process proceeds to step S3.

7. A P-LDPC code encoding device, characterized in that: The method comprises a memory and a processor, wherein the memory stores at least one program, and the at least one program is executed by the processor to implement the P-LDPC code encoding method according to any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that The storage medium stores at least one program, and the at least one program is executed by a processor to implement the P-LDPC code encoding method according to any one of claims 1 to 6.

9. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the encoding method of the P-LDPC code according to any one of claims 1 to 6 is implemented.

Images