Coding method and system based on deep learning and compressed sensing assistance

Abstract

The invention discloses a coding method and system based on deep learning and compressed sensing assistance, and belongs to the technical field of optical transmission. The method comprises the following steps: carrying out coding mapping on bit data, carrying out codebook design by using an AlexNet model, and training each codebook through five convolutional layers to obtain an error rate in a current environment; performing data mapping on the codebook with the lowest bit error rate; performing compressed sensing on the data signal mapped to the codebook; the method comprises the following steps of: firstly carrying out OFDM (Orthogonal Frequency Division Multiplexing) modulation on a signal subjected to compressed sensing, then carrying out signal transmission and signal demodulation, reconstructing the signal, demapping the reconstructed signal into a codebook signal, and finally decoding the codebook signal into an original signal through a decoder. The SCMA modulation is combined with the convolutional neural network technology and the compressed sensing technology, an encoder and a decoder of the SCMA are established, the bit error rate is minimized, signals are processed by using the compressed sensing theory, the amount of data needing to be collected, transmitted and processed is greatly reduced, and data are transmitted by occupying less subcarriers.

Description

Technical Field

[0001] The present invention belongs to the field of optical transmission technology, and specifically relates to a coding method and system based on deep learning and compressed sensing assistance. Background Art

[0002] 5G access is still based on the orthogonal frequency division multiplexing (OFDM) waveform, and has non-orthogonal waveforms and multiple access capabilities. Non-orthogonal multiple access methods effectively support higher capacity with greater flexibility and robustness, as well as adaptability to a large number of connections. These properties help provide a better user experience for enhanced mobile broadband (eMBB) services, and can even be mixed with other business types and other use cases identified for 5G. Regarding non-orthogonal multiple access schemes, there is a technology called sparse code multiple access (SCMA). With the development of non-orthogonal multiple access technology, sparse code multiple access technology (SCMA) has been increasingly used. In order to meet the needs of 5G communications, its innovative use of codebook technology at the transmitting end combines the modulation and spread spectrum technologies in traditional communications, increasing the flexibility of user selection and resource utilization. The performance of SCMA depends largely on the design of the codebook. The codebook mapping of SCMA is similar to the layer mapping of long-term evolution (LET), that is, one or more SCMA layers are assigned to users / data streams, but at each SCMA layer, a mapping from information bits to codewords will also be performed, that is, the SCMA modulator maps the input bits to complex multidimensional codewords selected from the SCMA codebook of a specific layer. SCMA codewords are sparse, that is, only a few of their entries are non-zero, and the rest are zero. Information bits are assigned to multidimensional complex codewords using predefined codebooks, which are not orthogonal to each other. The received signal can be decoded using maximum likelihood (ML) decoding or by message passing algorithm (MPA). However, there are loopholes in designing codebooks in an artificial way, because the codewords in the codebook are not orthogonal to each other and consist of multidimensional eigenvalues. Different codebooks need to be constructed in different environments, which leads to too high requirements on algorithm efficiency. In addition, decoding uses ML or iterative MPA, both of which have high computational overheads, limiting the real-time performance of SCMA. Therefore, it is of great significance to optimize the codebook design. Summary of the invention

[0003] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a coding method and system based on deep learning and compressed sensing assistance, which solves the problems in the prior art.

[0004] The purpose of the present invention can be achieved through the following technical solutions:

[0005] The encoding method based on deep learning and compressed sensing assistance includes the following steps:

[0006] Encode and map the bit data, use the AlexNet model to design the codebook, and each codebook is trained through five convolutional layers to obtain the bit error rate in the current environment;

[0007] Perform data mapping on the codebook with the lowest bit error rate;

[0008] Performing compressed sensing on the data signal mapped to the codebook;

[0009] The compressed sensing signal is first modulated by OFDM, then transmitted, and then demodulated. After that, the signal is reconstructed and demapped into a codebook signal, and finally the codebook signal is decoded into the original signal through a decoder.

[0010] Furthermore, the codebook design using the AlexNet model relies on CNN training to make the encoder f kj (r; θ f ), decoder g(y;θ g ) correctly reconstructs the original data r; the end-to-end loss function of the CNN considering both the encoder and the decoder is:

[0011]

[0012] In the formula, h is the channel vector, r is the target value, f is the encoder, g is the decoder, The data reconstructed by the CNN decoder is

[0013] Furthermore, the AlexNet model in CNN uses a dual-GPU-based convolution operation method and uses the Relu activation function instead of the T activation function. The Relu activation function is expressed as:

[0014] f(x)=max(0,x).

[0015] Furthermore, the data mapping is to transmit 6 user data using 4 subcarriers, and the mapping process is: 6 users all transmit two code elements, each code element occupies one subcarrier, and the 4 subcarriers are represented by b1, b2, b3 and b4 respectively; the first user data is mapped to b1 and b2, the second user is mapped to b1 and b3, the third user is mapped to b1 and b4, the fourth user is mapped to b2 and b3, the fifth user is mapped to b2 and b4, and the last one is mapped to b3 and b4.

[0016] Furthermore, performing compressed sensing on the data signal mapped to the codebook is sampling and compressing the signal, and the steps include:

[0017] S31, assuming that x is a one-dimensional discrete signal of length N, then a set of standard orthogonal basis The linear combination representation is used to form an N×N basis matrix by taking them as column vectors, i.e., Ψ = {Ψ1, Ψ2, …, Ψ N}, then the signal x can be represented by these basis vectors as follows:

[0018]

[0019] where θ = [θ1, θ2, …… θ N T is the sparse representation coefficient of the signal x in the basis Ψ, and only a small number of values in θ are large, while most values are zero or close to zero; if there are S non-zero elements among them,

[0020] the signal x is said to be S-sparse;

[0021] For S << N, a measurement matrix Φ of size M×N (M << N) is designed, and this matrix is uncorrelated with the transformation basis Ψ; the measurement of the signal x through the measurement matrix Φ can obtain an M×1 measurement vector:

[0022] y = Φx = ΦΨθ = Αθ

[0023] The high-dimensional signal Ψθ is mapped to a low-dimensional space under the measurement of the observation matrix Φ, that is, the compressed signal y is obtained, and the sampling compression of the signal is completed.

[0024] Furthermore, the steps to decode the codebook signal into the original signal are as follows:

[0025] S41, for each resource, calculate the probability of each possible codeword on this resource according to the received signal and channel gain, and pass these probabilities to the corresponding user nodes;

[0026] S42, each user node updates the probability of the codeword it selects according to the probability information received from other resource nodes, and passes these probabilities back to the resource nodes;

[0027] S43, repeat S41 - S42 until a predetermined number of iterations is reached;

[0028] S44, after the iteration ends, calculate the likelihood ratio of each bit according to the final probability distribution of the user nodes, and perform bit decision according to the likelihood ratio to recover the original signal.

[0029] A coding system based on deep learning and compressive sensing assistance includes:

[0030] Codebook design module: performs coding mapping on bit data, uses the AlexNet model for codebook design, and each codebook obtains the bit error rate in the current environment after being trained through five convolutional layers;

[0031] ​Data mapping module: performs data mapping on the codebook with the lowest bit error rate;

[0032] Compressed sensing module: performs compressed sensing on the data signal mapped to the codebook;

[0033] And, reconstruction decoding module: the compressed sensing signal is first modulated by OFDM, then transmitted, and then demodulated, and then the signal is reconstructed and demapped into a codebook signal, and finally the codebook signal is decoded into the original signal through a decoder.

[0034] A computer storage medium stores a readable program, which can execute the above-mentioned encoding method based on deep learning and compressed sensing assistance when the program is run.

[0035] An electronic device, comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other through the communication bus;

[0036] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the above-mentioned encoding method based on deep learning and compressed sensing assistance.

[0037] A computer program product includes computer instructions, wherein the computer instructions instruct a computing device to perform operations corresponding to the above-mentioned encoding method based on deep learning and compressed sensing assistance.

[0038] Beneficial effects of the present invention:

[0039] The present invention combines SCMA modulation with convolutional neural network technology and compressed sensing technology to establish an encoder and decoder for SCMA. They are composed of a basic AlexNet unit consisting of multiple convolutional layers and fully connected layers, in which the codebook design and signal encoding and decoding are performed using AlexNet; the proposed neural network is autonomously learned according to the environment, thereby minimizing the bit error rate; at the same time, the system uses compressed sensing theory to process signals, greatly reducing the amount of data that needs to be collected, transmitted, and processed, and realizing data transmission using fewer subcarriers. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0041] Figure 1 It is a flow chart of the encoding method assisted by deep learning and compressed sensing of the present invention;

[0042] Figure 2 It is the AlexNet network structure diagram of the present invention;

[0043] Figure 3 It is a comparison diagram of the codebook mapping process based on compressed sensing of the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] Example 1

[0046] like Figure 1 As shown, the encoding method based on deep learning and compressed sensing assistance includes the following steps:

[0047] S1, encodes and maps the bit data, uses the AlexNet model to design the codebook, and each codebook is trained through five convolutional layers to obtain the bit error rate in the current environment;

[0048] First, it is necessary to systematically collect and preprocess the relevant data of the SCMA system, including key information such as factor graph matrix, user data, and resource block allocation. Subsequently, a set of targeted feature extraction strategies is designed to extract features that are valuable for codebook design from complex SCMA signals. On this basis, an AlexNet deep learning model is constructed, which is adapted to the characteristics of the SCMA system due to its multi-layer convolution, pooling, and fully connected design ideas. It will be used to learn the inherent laws of the SCMA system and the complex relationship between features, thereby guiding the selection of codebook parameters, such as the configuration of the constellation diagram, the dimension and sparsity of the codebook, etc. After sufficient model training, it is used to predict and evaluate the performance of the designed codebook, focusing on key indicators such as bit error rate, throughput, and resource efficiency. According to the evaluation results, the codebook design, as well as the parameters and structure of the deep learning model, are continuously adjusted and optimized to form an iterative optimization closed loop.

[0049] The network structure of the AlexNet model is as follows Figure 2 As shown, the codebook design using the AlexNet model relies on convolutional neural network (CNN) training. The proposed CNN structure must be trained so that the encoder f kj (r; θ f ), decoder g(y;θ g) can correctly reconstruct the original data r; therefore, the end-to-end loss function of CNN considering both the encoder and the decoder can be written as:

[0050]

[0051] In the formula, h is the channel vector, r is the target value, f is the encoder, g is the decoder, The data reconstructed by the CNN decoder is

[0052] In this scheme, the input data symbol r is randomly generated during the training process and fed into the encoder. Then, h and n are randomly generated and added to the output of the encoder to obtain hf(r; θ f )+n, thereafter, hf(r;θ f +n) simulates the input of the receiver in the actual channel environment and is input into the decoder to reconstruct r. Note that the randomly generated r, h and n constitute the training data set of this scheme. Among them, the impact of noise n is represented by the damage level η, where η represents the average damage (noise) power divided by the average transmitted signal power (E[|x| 2 ]), that is, η=σ 2 / E[|x| 2 ]. In order to build the best encoder and decoder, the damage level must be correctly selected to meet the needs of multiple training of different codebooks to obtain the bit error rate. Each user has a completely different and compatible codebook, so more codebooks need to be trained, and a faster training speed can meet the needs.

[0053] The AlexNet model in the convolutional neural network uses a dual-GPU-based convolution operation method and uses the Relu activation function instead of the T activation function, which greatly improves the training speed and can solve the problems encountered in the training codebook process. The Relu activation function is expressed as follows:

[0054] f(x)=max(0,x)

[0055] After the user codebook is trained through five convolutional layers, the noise problem of the user data itself will be solved in the fully connected layer. By using the dropout technology, the overfitting of the neural network can be prevented more effectively. The overfitting of the neural network can be regarded as an overly complex model, which can easily exploit seemingly correct but actually useless relationships, which often leads to misjudgment in the end. Therefore, each time the user codebook is input, the neural network tries a different structure, and all these structures share weights. For a certain layer of neurons, some neurons are randomly deleted by a defined probability, while keeping the number of neurons in the input layer and the output layer unchanged, and then the parameters are updated according to the learning method of the neural network. In the next iteration, some neurons are randomly deleted again until the training is completed, ensuring the accuracy of the final bit error rate.

[0056] Applying the AlexNet model to the present invention can better construct a user codebook in high dimensions, and combining the high-dimensional user codebook with the AlexNet model can improve the system robustness while ensuring lower decoding complexity, thereby having better overall system performance.

[0057] S2, performing data mapping on the obtained codebook with the lowest bit error rate;

[0058] Such data mapping can be understood as transmitting 6 user data using 4 subcarriers. The specific mapping process is: 6 users transmit two symbols, each symbol occupies one subcarrier, and the 4 subcarriers are represented by b1, b2, b3 and b4; the first user data is mapped to b1 and b2, the second user is mapped to b1 and b3, the third user is mapped to b1 and b4, the fourth user is mapped to b2 and b3, the fifth user is mapped to b2 and b4, and the last user is mapped to b3 and b4. This ensures the uniqueness of each user codebook, reduces the occupancy of subcarriers, and achieves efficient use of spectrum.

[0059] S3, compresses the data signal mapped to the codebook to further optimize the data, reduce the use of subcarriers, and reduce complexity;

[0060] Compressed sensing of data signals mapped to the codebook is to sample and compress the signals. Different from the traditional signal acquisition process, signal sampling and compression are performed simultaneously in compressed sensing theory. This process discards the redundant information in the current signal sampling and directly obtains compressed samples from the continuous-time signal transformation; Figure 3 As shown, the step of sampling and compressing the data signal mapped to the codebook includes:

[0061] S31, assuming that x is a one-dimensional discrete signal of length N, then a set of standard orthogonal basis is represented by a linear combination and formed into an N×N basis matrix as column vectors, i.e., Ψ = {Ψ1, Ψ2, …, Ψ N}, then the signal x can be represented by these basis vectors as:

[0062]

[0063] where θ = [θ1, θ2, …… θ N T is the sparse representation coefficient of the signal x under the basis Ψ, and only a small part of the values in θ are large, while most of the values are zero or close to zero; if there are S non-zero elements among them, the signal x is said to be S-sparse;

[0064] For S << N, design a measurement matrix Φ of size M×N (M << N), which is uncorrelated with the transformation basis Ψ; the measurement of the signal x through the measurement matrix Φ can obtain an M×1 measurement vector:

[0065] y = Φx = ΦΨθ = Αθ

[0066] The high-dimensional signal Ψθ is mapped to a low-dimensional space under the measurement of the observation matrix Φ, that is, the compressed signal y is obtained, completing the sampling compression of the signal.

[0067] S4, the signal after compressive sensing first undergoes OFDM modulation, then signal transmission, followed by signal demodulation, and then the signal is reconstructed and demapped into a codebook signal, and finally the codebook signal is decoded into the original signal by the decoder;

[0068] During transmission, the codeword of each layer is selected based on the input bit sequence. After compressive sensing, the codewords from different layers are combined before the OFDM modulator and transmitted through a smaller number of subcarriers. The signal modulated by the OFDM modulator is transmitted through the channel. The problem of reconstructing the signal at the receiving end is actually to solve the above equation, that is, to solve a system of linear equations, and the original data can be obtained after demapping and passing through the decoder.

[0069] The process of signal OFDM modulation is as follows: convert the serial data into multiple parallel data streams, the data streams undergo codebook mapping to form multiple complex sub-symbols. Each complex sub-symbol corresponds to one path of data, the signal is sampled and observed by compressive sensing, and the signal is mapped to a low-dimensional space through the observation matrix, and then the data is allocated to different subcarriers. Perform an inverse fast Fourier transform (IFFT) on the complex sub-symbols on the codebook. The IFFT transforms the complex sub-symbols in the frequency domain into real-valued samples in the time domain, add a cyclic prefix before the real symbols, and finally perform digital-to-analog conversion and enter the channel for transmission.

[0070] ​The process of signal demodulation, reconstruction, and demapping is as follows: remove the cyclic prefix from the received signal, perform FFT processing on the symbol block after removing the cyclic prefix to obtain a frequency domain signal, perform frequency domain equalization on the signal, reconstruct the equalized data through the observation matrix, demap the reconstructed signal, and restore the obtained signal to a bit stream of the transmitted data through a decoder.

[0071] The steps to decode the codebook signal into the original signal are:

[0072] S41, for each resource (or subcarrier), the probability of each possible codeword on the resource is calculated according to the received signal and the channel gain, and the probabilities are transmitted to the corresponding user node;

[0073] S42, each user node updates the probability of the codeword selected by itself according to the probability information received from other resource nodes, and transmits these probabilities back to the resource node;

[0074] S43, repeating S41-S42 until a predetermined number of iterations is reached;

[0075] S44, after the iteration is completed, the likelihood ratio of each bit is calculated according to the final user node probability distribution, and the bit decision is performed according to the likelihood ratio to restore the original signal.

[0076] Based on similar inventive concepts, an embodiment of the present invention also provides a computer storage medium storing a readable program, which can execute the above-mentioned encoding method based on deep learning and compressed sensing assistance when the program is running.

[0077] Based on similar inventive concepts, an embodiment of the present invention provides an electronic device, comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other through the communication bus;

[0078] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the above-mentioned encoding method based on deep learning and compressed sensing assistance.

[0079] Based on similar inventive concepts, an embodiment of the present invention also provides a computer program product, including computer instructions, which instruct a computing device to perform operations corresponding to the above-mentioned encoding method based on deep learning and compressed sensing assistance.

[0080] Example 2

[0081] The coding system based on deep learning and compressed sensing includes:

[0082] Codebook design module: encodes and maps bit data, uses the AlexNet model to design codebooks, and obtains the bit error rate in the current environment after each codebook is trained through five convolutional layers;

[0083] Data mapping module: performs data mapping on the codebook with the lowest bit error rate;

[0084] Compressed sensing module: performs compressed sensing on the data signal mapped to the codebook;

[0085] Reconstruction and decoding module: The compressed sensing signal is first modulated by OFDM, then transmitted, and then demodulated. After that, the signal is reconstructed and demapped into a codebook signal, and finally the codebook signal is decoded into the original signal through a decoder.

[0086] The method of the present invention may be implemented in hardware, firmware, or as software or computer code that may be stored in a recording medium (such as a CDROM, RAM, floppy disk, hard disk, or magneto-optical disk), or as computer code that is originally stored in a remote recording medium or a non-temporary machine-readable medium downloaded over a network and will be stored in a local recording medium, so that the method described herein may be stored in such software processing on a recording medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware (such as an ASIC or FPGA). It is understood that a computer, processor, microprocessor controller, or programmable hardware includes a storage component (e.g., RAM, ROM, flash memory, etc.) that can store or receive software or computer code, and when the software or computer code is accessed and executed by a computer, processor, or hardware, the method described herein is implemented. In addition, when a general-purpose computer accesses the code for implementing the method shown herein, the execution of the code converts the general-purpose computer into a dedicated computer for executing the method shown herein.

[0087] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements all fall within the scope of the present invention to be protected.

Claims

1. A coding method based on deep learning and compressed sensing, characterized in that: It includes the following steps: Perform coding mapping on bit data, design codebooks using the AlexNet model, and obtain the bit error rate in the current environment after each codebook is trained through five convolutional layers; Perform data mapping on the codebook with the lowest bit error rate; Perform compressive sensing on the data signal mapped to the codebook; The signal after compressive sensing first undergoes OFDM modulation, then signal transmission, followed by signal demodulation. After that, the signal is reconstructed and demapped into a codebook signal, and finally, the codebook signal is decoded into the original signal by a decoder.

2. The encoding method based on deep learning and compressed sensing assistance according to claim 1, characterized in that: The codebook design using the AlexNet model relies on CNN training to make the encoder f kj (r; θ f ), decoder g(y;θ g ) correctly reconstructs the original data r; the end-to-end loss function of the CNN considering both the encoder and the decoder is: In the formula, h is the channel vector, r is the target value, f is the encoder, g is the decoder, The data reconstructed by the CNN decoder is 3. The encoding method based on deep learning and compressed sensing assistance according to claim 2, characterized in that: The AlexNet model in CNN uses a convolutional operation method based on dual GPUs, and simultaneously replaces the T activation function with the Relu activation function. The Relu activation function is expressed as: f(x) = max(0, x).

4. The encoding method based on deep learning and compressed sensing assistance according to claim 1, characterized in that: The data mapping is to transmit 6 user data using 4 subcarriers. The mapping process is as follows: 6 users all transmit two symbols, each symbol occupies one subcarrier, and the 4 subcarriers are represented by b1, b2, b3, and b4 respectively; map the first user data to b1 and b2, the second user to b1 and b3, the third user to b1 and b4, the fourth user to b2 and b3, the fifth user to b2 and b4, and the last one to b3 and b4.

5. The encoding method based on deep learning and compressed sensing assistance according to claim 1, characterized in that: Performing compressive sensing on the data signal mapped to the codebook is to sample and compress the signal. The steps include: S31, assuming that x is a one-dimensional discrete signal of length N, then a set of standard orthogonal basis The linear combination of is represented as a column vector to form an N×N Wiki matrix, that is, Ψ={Ψ1,Ψ2,…,Ψ N }, then the signal x can be expressed by the basis vector as: Among them, θ = [θ1, θ2,…θ N ] T is the sparse representation coefficient of the signal x in the basis Ψ, and only a small part of θ has a large value, and most of the values ​​are zero or close to zero; if there are S non-zero elements, the signal x is called S-sparse; S32, design a measurement matrix Φ of size M×N (M << N), which is uncorrelated with the transform basis Ψ; the signal x can be measured by the measurement matrix Φ to obtain an M×1 measurement vector: y = Φx = ΦΨθ = Αθ The high-dimensional signal Ψθ is mapped to a low-dimensional space under the measurement of the observation matrix Φ, that is, the compressed signal y is obtained, and the sampling and compression of the signal are completed.

6. The encoding method based on deep learning and compressed sensing assistance according to claim 1, characterized in that: The steps to decode the codebook signal into the original signal are: S41, for each resource, calculate the probability of each possible codeword on this resource according to the received signal and channel gain, and transfer these probabilities to the corresponding user node; S42, each user node updates the probability of the codeword it selects according to the probability information received from other resource nodes, and returns these probabilities to the resource node; S43, repeat S41 - S42 until a predetermined number of iterations is reached; S44, after the iteration ends, calculate the likelihood ratio of each bit according to the final probability distribution of the user nodes, and perform bit decision according to the likelihood ratio to recover the original signal.

7. A coding system based on deep learning and compressed sensing, characterized in that: It includes: Codebook design module: Perform coding mapping on bit data, design codebooks using the AlexNet model, and obtain the bit error rate in the current environment after each codebook is trained through five convolutional layers; Data mapping module: Perform data mapping on the codebook with the lowest bit error rate; Compressive sensing module: Perform compressive sensing on the data signal mapped to the codebook; And a reconstruction and decoding module: The signal after compressive sensing first undergoes OFDM modulation, then signal transmission, followed by signal demodulation. After that, the signal is reconstructed and demapped into a codebook signal, and finally, the codebook signal is decoded into the original signal by a decoder.

8. A computer storage medium storing a readable program, characterized in that: When the program is running, it can execute the encoding method based on deep learning and compressed sensing assistance as described in any one of claims 1 to 6.

9. An electronic device, characterized in that: include: A processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the encoding method based on deep learning and compressed sensing assistance as described in any one of claims 1-6.

10. A computer program product comprising computer instructions, characterized in that: The computer instructions instruct the computing device to perform operations corresponding to the encoding method based on deep learning and compressed sensing assistance as described in any one of claims 1-6.

Images