CN112232129A - Electromagnetic information leakage signal simulation system and method based on generation countermeasure network - Google Patents
Electromagnetic information leakage signal simulation system and method based on generation countermeasure network Download PDFInfo
- Publication number
- CN112232129A CN112232129A CN202010981486.7A CN202010981486A CN112232129A CN 112232129 A CN112232129 A CN 112232129A CN 202010981486 A CN202010981486 A CN 202010981486A CN 112232129 A CN112232129 A CN 112232129A
- Authority
- CN
- China
- Prior art keywords
- electromagnetic
- signal
- information leakage
- display screen
- image information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000004088 simulation Methods 0.000 title claims abstract description 59
- 238000000034 method Methods 0.000 title claims abstract description 40
- 230000006870 function Effects 0.000 claims abstract description 12
- 238000005457 optimization Methods 0.000 claims abstract description 5
- 239000000523 sample Substances 0.000 claims description 60
- 238000012549 training Methods 0.000 claims description 23
- 230000008569 process Effects 0.000 claims description 20
- 239000013598 vector Substances 0.000 claims description 18
- 238000010606 normalization Methods 0.000 claims description 14
- 238000013507 mapping Methods 0.000 claims description 11
- 230000004913 activation Effects 0.000 claims description 10
- 238000004422 calculation algorithm Methods 0.000 claims description 3
- 238000013527 convolutional neural network Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 4
- 238000005516 engineering process Methods 0.000 abstract description 3
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 238000013135 deep learning Methods 0.000 abstract description 2
- 238000004364 calculation method Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 7
- 238000013528 artificial neural network Methods 0.000 description 6
- 238000013461 design Methods 0.000 description 4
- 230000007246 mechanism Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000005670 electromagnetic radiation Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000016273 neuron death Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Signal Processing (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The electromagnetic information leakage signal simulation system based on the generation countermeasure network comprises a signal acquisition device, a display screen image information leakage discriminator and an electromagnetic signal generator; according to the invention, from the angle of deep learning, the electromagnetic signal of the analog display screen is learned by using a method for generating the confrontation network, the optimization is further improved after feature fitting comparison with the original electromagnetic signal is carried out, and after iterative updating is completed, analog simulation signal data is continuously perfected, so that the analog effect of the electromagnetic signal can be realized to a higher degree. The invention aims to solve the problems of electromagnetic information leakage and the like by using related technologies with good artificial intelligence while the difficulty is faced by the traditional reconstruction technical means, so that the electronic information equipment and the method cooperate with each other to provide the safety protection function to the maximum extent.
Description
Technical Field
The invention relates to the technical field of simulation, in particular to an electromagnetic information leakage signal simulation system and method based on a generation countermeasure network.
Background
As shown in fig. 1, the display screen of the electronic information device emits electromagnetic waves to the external environment during operation, thereby generating electromagnetic radiation for unintentional, non-subjective communication. These electromagnetic radiation signals are called electromagnetic leakage signals, which often contain image information on the display screen, resulting in information leakage of the displayed image, threatening the information security of the device.
In the research of the electromagnetic information leakage problem, the reconstruction technology of the electromagnetic signal is developed. The mechanism mainly comprises the steps of reconnaissance analysis, parameter measurement, signal reconstruction and the like, namely after the electromagnetic leakage signal is obtained, the key information in the electromagnetic leakage signal is further obtained by performing parameter measurement and characteristic analysis on the signal, so that the reconstruction of the original information on the equipment is realized. To a certain extent, the reconstructed signals have high similarity with the intercepted signals, and a deceptive electromagnetic camouflage protection effect is achieved. However, the electromagnetic signal environment is complex and changeable, and the conventional signal reconstruction mechanism is difficult to acquire various types of signal key information and accurately reconstruct the electromagnetic signal in the complex electromagnetic environment.
Disclosure of Invention
The invention provides an electromagnetic information leakage signal simulation system and method based on a generation countermeasure network, and mainly aims to solve the problem that key information of various signals is difficult to obtain and electromagnetic signals are accurately reconstructed in a complex electromagnetic environment by a traditional signal reconstruction mechanism.
In order to solve the technical problems, the invention adopts the following technical scheme:
the electromagnetic information leakage signal simulation system based on the generation countermeasure network comprises a signal acquisition device, a display screen image information leakage discriminator and an electromagnetic signal generator;
the signal acquisition device is used for sensing electromagnetic signals containing image information from a data line of the display screen, and receiving and storing the electromagnetic signals acquired by the signal probe into an original electromagnetic signal sample through the signal receiver;
the display screen image information leakage discriminator is used for extracting characteristics from an original electromagnetic signal sample to be used as a discrimination basis of the display screen image information leakage discriminator, finally forming classification discrimination of electromagnetic information leakage, guiding the electromagnetic signal generator to train, and evaluating whether data input is a real signal or a generated electromagnetic simulation signal by analyzing the original electromagnetic signal sample and electromagnetic simulation signal sample data from the electromagnetic signal generator;
the electromagnetic signal generator is used for generating an electromagnetic simulation signal containing image information by random noise, learning characteristic distribution in original electromagnetic signal sample data, directly generating a simulation electromagnetic signal conforming to the distribution rule in the learning process, and receiving loss function feedback from the display screen image information leakage discriminator after evaluating the generated sample.
Furthermore, the display screen image information leakage discriminator comprises four convolution layers and a full connection layer, electromagnetic signals sequentially pass through the four convolution layers, and discrimination results are output from the full connection layer.
Furthermore, the convolutional layer comprises a one-dimensional convolution kernel, a Batch normalization layer Batch Norm and an LReLU activation layer in sequence for the electromagnetic signal.
Furthermore, the electromagnetic signal generator comprises a feature mapping layer and a plurality of deconvolution layers, random noise is subjected to dimension-raising mapping through the feature mapping layer to be an initial generation vector, the initial generation vector sequentially passes through the plurality of deconvolution layers, and the electromagnetic simulation signal is generated by the last deconvolution layer.
Furthermore, the first three deconvolution layers of the four deconvolution layers respectively comprise a one-dimensional deconvolution kernel, a Batch normalization layer Batch Norm and a ReLU activation layer, which are sequentially formed by the generated vectors, and the last convolution layer comprises a one-dimensional deconvolution kernel, a Batch normalization layer Batch Norm and a tanh activation layer, which are sequentially formed by the generated vectors.
The electromagnetic information leakage signal simulation method based on the generation countermeasure network comprises the following steps:
s1, constructing a display screen image information leakage discriminator and an electromagnetic signal generator respectively according to the characteristics of electromagnetic information leakage signals based on a convolutional neural network method;
s2, sensing an electromagnetic signal containing image information from a data line of a display screen by using a signal acquisition device, and receiving and storing the electromagnetic signal acquired by the signal acquisition device into an original electromagnetic signal sample through a signal receiver;
s3, the display screen image information leakage discriminator extracts characteristics from an original electromagnetic signal sample to be used as a discrimination basis of the display screen image information leakage discriminator, and finally classification discrimination of electromagnetic information leakage is formed;
s4, inputting the random noise into an electromagnetic signal generator, and generating an electromagnetic simulation signal containing image information by the electromagnetic signal generator;
s5, training a display screen image information leakage discriminator, wherein the display screen image information leakage discriminator guides an electromagnetic signal generator to train, and judges whether the data input is a real signal or a generated electromagnetic simulation signal by analyzing an original electromagnetic signal sample and electromagnetic simulation signal sample data from the electromagnetic signal generator;
s6, training an electromagnetic signal generator, wherein the electromagnetic signal generator learns the characteristic distribution in original electromagnetic signal sample data, directly generates a simulation electromagnetic signal according with the distribution rule in the learning process, receives the loss function feedback from a display screen image information leakage discriminator after evaluating a generated sample, and further updates and perfects the electromagnetic signal generator, so that an electromagnetic simulation signal closer to real sample data can be generated;
s7, in the training process, the display screen image information leakage discriminator and the electromagnetic signal generator continuously compete, so that the discrimination capability and the generation capability of each are improved, the generation of the countermeasure network is continuously updated in an iterative mode through the countermeasure competition between the display screen image information leakage discriminator and the electromagnetic signal generator, and finally the electromagnetic simulation signal which cannot be identified by the display screen image information leakage discriminator is obtained.
Further, the signal acquisition device in step S2 adopts a clamp type electromagnetic signal probe.
Furthermore, the training of the electromagnetic signal generator adopts an alternate optimization mode, the iteration process is divided into two stages, the display screen image information leakage discriminator is firstly kept in the first stage, and the electromagnetic signal generator is optimized, so that the probability that the generated electromagnetic simulation signal is evaluated as a real signal by the display screen image information leakage discriminator is as high as possible; and in the second stage, the electromagnetic signal generator is kept, the discrimination model of the display screen image information leakage discriminator is optimized, and the classification accuracy of the display screen image information leakage discriminator is improved.
Further, in the process of generating the countermeasure network for training, the basic working algorithm of the display screen image information leakage discriminator part is as follows: in the iterative training process of a plurality of times, a small batch of generated samples are selected from the sample content generated by the electromagnetic signal generator in each cycle, meanwhile, a small batch of real samples are selected from the real sample data set, the small batch of real samples and the real sample data set are integrated into an input sample, and the discriminator is updated by continuously increasing the random gradient.
As can be seen from the above description of the present invention, compared with the prior art, the present invention has the following advantages: according to the invention, from the angle of deep learning, the electromagnetic signal of the analog display screen is learned by using a method for generating the confrontation network, the optimization is further improved after feature fitting comparison with the original electromagnetic signal is carried out, and after iterative updating is completed, analog simulation signal data is continuously perfected, so that the analog effect of the electromagnetic signal can be realized to a higher degree. The invention aims to solve the problems of electromagnetic information leakage and the like by using related technologies with good artificial intelligence while the difficulty is faced by the traditional reconstruction technical means, so that the electronic information equipment and the method cooperate with each other to provide the safety protection function to the maximum extent.
Drawings
FIG. 1 is a schematic diagram of electromagnetic leakage of a display screen image.
Fig. 2 is a schematic block diagram of the present invention.
FIG. 3 is a schematic diagram of the image information leakage discriminator of the display screen according to the present invention.
FIG. 4 is a schematic diagram of a convolutional layer of the present invention.
Fig. 5 is a schematic diagram of an electromagnetic signal generator of the present invention.
FIG. 6 is a schematic diagram of the first three deconvolution layers in the electromagnetic signal generator of the present invention.
FIG. 7 is a schematic diagram of the last deconvolution layer in the electromagnetic signal generator of the present invention.
Detailed Description
Referring to fig. 2, the electromagnetic information leakage signal simulation system based on the generation countermeasure network comprises a signal acquisition device 1, a display screen image information leakage discriminator 3 and an electromagnetic signal generator 3;
the signal acquisition device 1 is used for sensing an electromagnetic signal containing image information from a data line of the display screen, and receiving and storing the electromagnetic signal acquired by the signal probe into an original electromagnetic signal sample through the signal receiver;
the display screen image information leakage discriminator 2 is used for extracting characteristics from an original electromagnetic signal sample, taking the characteristics as a discrimination basis of the display screen image information leakage discriminator 2, finally forming classification discrimination of electromagnetic information leakage, guiding the electromagnetic signal generator 3 to train, and evaluating whether the data input is a real signal or a generated electromagnetic simulation signal by analyzing the original electromagnetic signal sample and the electromagnetic simulation signal sample data from the electromagnetic signal generator 3;
and the electromagnetic signal generator 3 is used for generating an electromagnetic simulation signal containing image information by using random noise, learning characteristic distribution in original electromagnetic signal sample data, directly generating a simulation electromagnetic signal conforming to the distribution rule in the learning process, and receiving loss function feedback from the display screen image information leakage discriminator 2 after evaluating the generated sample.
Referring to fig. 2,3 and 4, the display screen image information leakage discriminator 2 includes four convolutional layers 21 and one full-link layer 22, and electromagnetic signals sequentially pass through the four convolutional layers 21 and the discrimination result is output from the full-link layer 22. The convolutional layer 21 includes a one-dimensional convolutional kernel 211, a Batch normalization layer Batch Norm212, and an lreol active layer 213, in this order, for electromagnetic signals. In the design of the network structure of the display screen image information leakage discriminator 2, a neural network is constructed by using one-dimensional convolution layers, and the parameter settings of each layer are shown in table 1
Table 1 display screen image information leakage discriminator network parameters
The four convolution layers 21 of the display screen image information leakage discriminator 2 all adopt one-dimensional convolution kernels 211 with the same size, the length of the convolution kernels is 64, and the sliding step length in the convolution calculation process is 4. The function of the one-dimensional convolution kernel 211 is to extract characteristics from the original electromagnetic signal, which is used as a criterion of the display screen image information leakage discriminator 2, and finally forms classification discrimination of electromagnetic information leakage by extracting features layer by layer of the four convolution layers 21.
The convolution is calculated as follows:
where L is the index number of each layer in the neural network, X(L)And X(L+1)Respectively input and output feature vectors, W, of the L-th layer in the calculation process(L)Is the convolution kernel vector of the L-th neural network, B(L)Is the bias vector for layer L.
According to the method, the Batch normalization layer Batch Norm212 is added behind the one-dimensional convolution kernel 211, the electromagnetic signal characteristic data extracted by the one-dimensional convolution kernel 211 is subjected to offset and scale scaling adjustment, and a higher learning rate can be used, so that the network training speed is increased, the generalization capability is increased to a certain extent, and the precision of the display screen image information leakage discriminator 2 is improved.
Batch normalization layer Batch Norm212 was calculated as follows:
first, the mean value mu of the batch sample is calculatedB
Wherein N is the number of samples input in a batch during neural network training, XiIs the ith electromagnetic signal sample.
The variance σ of the batch samples is then calculated2 B
Normalized calculation
In which epsilon is a constant to compensate for sigma2 BToo small to have an effect.
Scaling and offsetting
Yi＝γX'i+β (5)
Where γ is the scaling dimension, β is the offset, YiIs the output of Batch normalization layer Batch Norm 212.
the LReLU is a variant of the ReLU, sparsity of the ReLU is reduced to a certain degree in the display screen image information leakage discriminator 2, the problem of neuron death caused by the ReLU is relieved, and discrimination accuracy is improved.
Referring to fig. 2, 5, 6 and 7, the electromagnetic signal generator 3 includes a feature mapping layer 31 and a plurality of deconvolution layers 32, random noise is subjected to dimension-up mapping by the feature mapping layer 31 to be an initially generated vector, the initially generated vector passes through the plurality of deconvolution layers 32 in sequence, and an electromagnetic simulation signal is generated by the last deconvolution layer 32. The first three deconvolution layers 32 of the four deconvolution layers 32 each include a one-dimensional deconvolution kernel 321, a Batch normalization layer Batch Norm322, and a ReLU active layer 323, in which generated vectors sequentially pass through, and the last convolution layer 32 includes a one-dimensional deconvolution kernel 321, a Batch normalization layer Batch Norm322, and a tanh active layer 324, in which generated vectors sequentially pass through.
For the network structure design of the electromagnetic signal generator, a one-dimensional deconvolution is used for constructing a neural network, and in addition, a proper electromagnetic signal is selected as an activation function of input data according to the characteristic correlation of the electromagnetic signal to generate an analog electromagnetic signal. Random noise is used as the input of the electromagnetic signal generator 3, the feature mapping layer 31 performs up-dimensional mapping on the random noise to form an initial generation vector, and electromagnetic simulation signals are generated layer by layer. The parameter settings of the various layers of the electromagnetic signal generator are shown in Table 2
TABLE 2 electromagnetic Signal Generator network parameters
In the one-dimensional deconvolution kernel 321 adopted in the present invention, the one-dimensional deconvolution kernel has a size of 64 and a step size of 4. When calculating, firstly, the input signal is calculated according to the step length as 1: the ratio of 4 is filled up and filled with a value of 0. Example (c): inputting [1,2,3,4, …, N ], filling and amplifying, outputting [1,0,0,0,2,0,0,0,3,0,0,0,4,0,0,0, …,0,0, N ], and performing convolution calculation on the filling and amplifying signal according to a one-dimensional deconvolution kernel 321 to obtain a generated vector.
The calculation of Batch normalization layer Batch Norm322 is consistent with the calculation of Batch normalization layer Batch Norm212 for the convolutional layer.
ReLU is selected as the activation function for the first three deconvolution layers 32, and the vector Z is generatediEach element z inijThe calculation is as follows:
referring to fig. 2, the electromagnetic information leakage signal simulation method based on the generation countermeasure network comprises the following steps:
s1, constructing a display screen image information leakage discriminator and an electromagnetic signal generator respectively according to the characteristics of electromagnetic information leakage signals based on a convolutional neural network method;
s2, sensing an electromagnetic signal containing image information from a data line of a display screen by using a signal acquisition device, and receiving and storing the electromagnetic signal acquired by the signal acquisition device into an original electromagnetic signal sample through a signal receiver; the signal acquisition device adopts a caliper type electromagnetic signal probe.
S3, the display screen image information leakage discriminator extracts characteristics from an original electromagnetic signal sample to be used as a discrimination basis of the display screen image information leakage discriminator, and finally classification discrimination of electromagnetic information leakage is formed;
s4, inputting the random noise into an electromagnetic signal generator, and generating an electromagnetic simulation signal containing image information by the electromagnetic signal generator;
s5, training a display screen image information leakage discriminator, wherein the display screen image information leakage discriminator guides an electromagnetic signal generator to train, and judges whether the data input is a real signal or a generated electromagnetic simulation signal by analyzing an original electromagnetic signal sample and electromagnetic simulation signal sample data from the electromagnetic signal generator;
s6, training an electromagnetic signal generator, wherein the electromagnetic signal generator learns the characteristic distribution in original electromagnetic signal sample data, directly generates a simulation electromagnetic signal according with the distribution rule in the learning process, receives the loss function feedback from a display screen image information leakage discriminator after evaluating a generated sample, and further updates and perfects the electromagnetic signal generator, so that an electromagnetic simulation signal closer to real sample data can be generated;
s7, in the training process, the display screen image information leakage discriminator and the electromagnetic signal generator continuously compete, so that the discrimination capability and the generation capability of each are improved, the generation of the countermeasure network is continuously updated in an iterative mode through the countermeasure competition between the display screen image information leakage discriminator and the electromagnetic signal generator, and finally the electromagnetic simulation signal which cannot be identified by the display screen image information leakage discriminator is obtained.
The training of the electromagnetic signal generator adopts an alternate optimization mode, the iteration process is divided into two stages, in the first stage, the display screen image information leakage discriminator is firstly kept, and the electromagnetic signal generator is optimized, so that the probability that the generated electromagnetic simulation signal is evaluated as a real signal by the display screen image information leakage discriminator is as high as possible; and in the second stage, the electromagnetic signal generator is kept, the discrimination model of the display screen image information leakage discriminator is optimized, and the classification accuracy of the display screen image information leakage discriminator is improved.
In the process of generating the countermeasure network for training, the basic working algorithm of the display screen image information leakage discriminator part is as follows: in the iterative training process of a plurality of times, a small batch of generated samples are selected from the sample content generated by the electromagnetic signal generator in each cycle, meanwhile, a small batch of real samples are selected from the real sample data set, the small batch of real samples and the real sample data set are integrated into an input sample, the discriminator is updated by continuously improving the random gradient, and the definition formula is as follows
The first half of the above summation equation refers to a calculated value output after real data x is input to the display screen image information leakage discriminator model D (), where it is desirable to be as large as possible during training; the second half is a calculated value obtained by putting data z generated by the electromagnetic signal generator G () into the display screen image information leakage discriminator, and it is desirable that the smaller the value is, the better the value is, in the training, so that the overall factor of the second half can be made as large as possible.
The invention is based on a convolution neural network method, and respectively constructs a display screen image information leakage discriminator and an electromagnetic signal generator according to the characteristics of electromagnetic leakage signals. Training a display screen image information leakage discriminator by using an original display screen electromagnetic leakage signal sample; generating random noise into a simulation electromagnetic signal through an electromagnetic signal generator; in the continuous countermeasure of the display screen image information leakage discriminator and the electromagnetic signal generator, the two are continuously updated in an iterative manner, and finally the analog simulation signal which accords with the electromagnetic information leakage signal characteristic is obtained. The electromagnetic signal generated by the invention is brand new and is different from the existing electromagnetic leakage signal; the image information in the generated signal is also completely new and different from the original image already present.
The above description is only an embodiment of the present invention, but the design concept of the present invention is not limited thereto, and any insubstantial modifications made by using the design concept should fall within the scope of infringing the present invention.
Claims (9)
1. Electromagnetic information leakage signal simulation system based on generation countermeasure network, characterized by: the device comprises a signal acquisition device, a display screen image information leakage discriminator and an electromagnetic signal generator;
the signal acquisition device is used for sensing electromagnetic signals containing image information from a data line of the display screen, and receiving and storing the electromagnetic signals acquired by the signal probe into an original electromagnetic signal sample through the signal receiver;
the display screen image information leakage discriminator is used for extracting characteristics from an original electromagnetic signal sample to be used as a discrimination basis of the display screen image information leakage discriminator, finally forming classification discrimination of electromagnetic information leakage, guiding the electromagnetic signal generator to train, and evaluating whether data input is a real signal or a generated electromagnetic simulation signal by analyzing the original electromagnetic signal sample and electromagnetic simulation signal sample data from the electromagnetic signal generator;
the electromagnetic signal generator is used for generating an electromagnetic simulation signal containing image information by random noise, learning characteristic distribution in original electromagnetic signal sample data, directly generating a simulation electromagnetic signal conforming to the distribution rule in the learning process, and receiving loss function feedback from the display screen image information leakage discriminator after evaluating the generated sample.
2. The electromagnetic information leakage signal simulation system based on generation of countermeasure networks of claim 1, wherein: the display screen image information leakage discriminator comprises four convolution layers and a full connection layer, wherein electromagnetic signals sequentially pass through the four convolution layers, and discrimination results are output from the full connection layer.
3. The electromagnetic information leakage signal simulation system based on generation of countermeasure networks of claim 2, wherein: the convolution layer comprises a one-dimensional convolution kernel, a Batch normalization layer Batch Norm and an LReLU activation layer which are sequentially passed by electromagnetic signals.
4. The electromagnetic information leakage signal simulation system based on generation of countermeasure networks of claim 1, wherein: the electromagnetic signal generator comprises a feature mapping layer and four deconvolution layers, random noise is subjected to dimension-raising mapping through the feature mapping layer to form an initial generation vector, the initial generation vector sequentially passes through the four deconvolution layers, and an electromagnetic simulation signal is generated by the last deconvolution layer.
5. The electromagnetic information leakage signal simulation system based on generation of countermeasure networks of claim 4, wherein: the first three deconvolution layers of the four deconvolution layers respectively comprise a one-dimensional deconvolution kernel, a Batch normalization layer Batch Norm and a ReLU activation layer, which are sequentially formed by vectors, and the last convolution layer comprises a one-dimensional deconvolution kernel, a Batch normalization layer Batch Norm and a tanh activation layer, which are sequentially formed by vectors.
6. The method for simulating an electromagnetic information leakage signal based on a generative countermeasure network as claimed in any one of claims 1 to 5, wherein: the method comprises the following steps:
s1, constructing a display screen image information leakage discriminator and an electromagnetic signal generator respectively according to the characteristics of electromagnetic information leakage signals based on a convolutional neural network method;
s2, sensing an electromagnetic signal containing image information from a data line of a display screen by using a signal acquisition device, and receiving and storing the electromagnetic signal acquired by the signal acquisition device into an original electromagnetic signal sample through a signal receiver;
s3, the display screen image information leakage discriminator extracts characteristics from an original electromagnetic signal sample to be used as a discrimination basis of the display screen image information leakage discriminator, and finally classification discrimination of electromagnetic information leakage is formed;
s4, inputting the random noise into an electromagnetic signal generator, and generating an electromagnetic simulation signal containing image information by the electromagnetic signal generator;
s5, training a display screen image information leakage discriminator, wherein the display screen image information leakage discriminator guides an electromagnetic signal generator to train, and judges whether the data input is a real signal or a generated electromagnetic simulation signal by analyzing an original electromagnetic signal sample and electromagnetic simulation signal sample data from the electromagnetic signal generator;
s6, training an electromagnetic signal generator, wherein the electromagnetic signal generator learns the probability distribution of original electromagnetic signal sample data, directly generates a simulation electromagnetic signal according with the distribution rule in the learning process, receives the loss function feedback from a display screen image information leakage discriminator after evaluating a generated sample, and further updates and perfects the electromagnetic signal generator, so that an electromagnetic simulation signal closer to real sample data can be generated;
s7, in the training process, the display screen image information leakage discriminator and the electromagnetic signal generator continuously compete, so that the discrimination capability and the generation capability of each are improved, the generation of the countermeasure network is continuously updated in an iterative mode through the countermeasure competition between the display screen image information leakage discriminator and the electromagnetic signal generator, and finally the electromagnetic simulation signal which cannot be identified by the display screen image information leakage discriminator is obtained.
7. The method for electromagnetic information leakage signal simulation based on generation of countermeasure networks of claim 6, wherein: the signal acquisition device in the step S2 adopts a caliper type electromagnetic signal probe.
8. The method for electromagnetic information leakage signal simulation based on generation of countermeasure networks of claim 6, wherein: the training of the electromagnetic signal generator adopts an alternate optimization mode, the iteration process is divided into two stages, in the first stage, the display screen image information leakage discriminator is firstly kept, and the electromagnetic signal generator is optimized, so that the probability that the generated electromagnetic simulation signal is evaluated as a real signal by the display screen image information leakage discriminator is as high as possible; and in the second stage, the electromagnetic signal generator is kept, the discrimination model of the display screen image information leakage discriminator is optimized, and the classification accuracy of the display screen image information leakage discriminator is improved.
9. The method for electromagnetic information leakage signal simulation based on generation of countermeasure networks of claim 6, wherein: in the process of generating the countermeasure network for training, the basic working algorithm of the display screen image information leakage discriminator part is as follows: in the iterative training process of a plurality of times, a small batch of generated samples are selected from the sample content generated by the electromagnetic signal generator in each cycle, meanwhile, a small batch of real samples are selected from the real sample data set, the small batch of real samples and the real sample data set are integrated into an input sample, and the discriminator is updated by continuously increasing the random gradient.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010981486.7A CN112232129A (en) | 2020-09-17 | 2020-09-17 | Electromagnetic information leakage signal simulation system and method based on generation countermeasure network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010981486.7A CN112232129A (en) | 2020-09-17 | 2020-09-17 | Electromagnetic information leakage signal simulation system and method based on generation countermeasure network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112232129A true CN112232129A (en) | 2021-01-15 |
Family
ID=74107217
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010981486.7A Pending CN112232129A (en) | 2020-09-17 | 2020-09-17 | Electromagnetic information leakage signal simulation system and method based on generation countermeasure network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112232129A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108680796A (en) * | 2018-05-17 | 2018-10-19 | 集美大学 | Electromagnetic information leakage detecting system and method for computer display |
| CN112818876A (en) * | 2021-02-04 | 2021-05-18 | 成都理工大学 | Electromagnetic signal extraction and processing method based on deep convolutional neural network |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108680796A (en) * | 2018-05-17 | 2018-10-19 | 集美大学 | Electromagnetic information leakage detecting system and method for computer display |
| CN109389080A (en) * | 2018-09-30 | 2019-02-26 | 西安电子科技大学 | Hyperspectral image classification method based on semi-supervised WGAN-GP |
| CN109962747A (en) * | 2019-03-13 | 2019-07-02 | 中国电子科技集团公司第二十九研究所 | The photon neural network self-adaptive processing device of wideband electromagnetic signal |
| CN110555811A (en) * | 2019-07-02 | 2019-12-10 | 五邑大学 | SAR image data enhancement method and device and storage medium |
| CN110598843A (en) * | 2019-07-23 | 2019-12-20 | 中国人民解放军63880部队 | Generation countermeasure network organization structure based on discriminator sharing and training method thereof |
| CN110598530A (en) * | 2019-07-30 | 2019-12-20 | 浙江工业大学 | Small sample radio signal enhanced identification method based on ACGAN |
| CN110969123A (en) * | 2019-12-02 | 2020-04-07 | 集美大学 | Electromagnetic information leakage detection method based on frequency domain, terminal equipment and storage medium |
| CN111337243A (en) * | 2020-02-27 | 2020-06-26 | 上海电力大学 | ACGAN-based wind turbine generator planet wheel gearbox fault diagnosis method |
| CN111428648A (en) * | 2020-03-26 | 2020-07-17 | 五邑大学 | Electroencephalogram signal generation network, method and storage medium |
| CN111568412A (en) * | 2020-04-03 | 2020-08-25 | 中山大学 | Method and device for reconstructing visual image by utilizing electroencephalogram signal |
-
2020
- 2020-09-17 CN CN202010981486.7A patent/CN112232129A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108680796A (en) * | 2018-05-17 | 2018-10-19 | 集美大学 | Electromagnetic information leakage detecting system and method for computer display |
| CN109389080A (en) * | 2018-09-30 | 2019-02-26 | 西安电子科技大学 | Hyperspectral image classification method based on semi-supervised WGAN-GP |
| CN109962747A (en) * | 2019-03-13 | 2019-07-02 | 中国电子科技集团公司第二十九研究所 | The photon neural network self-adaptive processing device of wideband electromagnetic signal |
| CN110555811A (en) * | 2019-07-02 | 2019-12-10 | 五邑大学 | SAR image data enhancement method and device and storage medium |
| CN110598843A (en) * | 2019-07-23 | 2019-12-20 | 中国人民解放军63880部队 | Generation countermeasure network organization structure based on discriminator sharing and training method thereof |
| CN110598530A (en) * | 2019-07-30 | 2019-12-20 | 浙江工业大学 | Small sample radio signal enhanced identification method based on ACGAN |
| CN110969123A (en) * | 2019-12-02 | 2020-04-07 | 集美大学 | Electromagnetic information leakage detection method based on frequency domain, terminal equipment and storage medium |
| CN111337243A (en) * | 2020-02-27 | 2020-06-26 | 上海电力大学 | ACGAN-based wind turbine generator planet wheel gearbox fault diagnosis method |
| CN111428648A (en) * | 2020-03-26 | 2020-07-17 | 五邑大学 | Electroencephalogram signal generation network, method and storage medium |
| CN111568412A (en) * | 2020-04-03 | 2020-08-25 | 中山大学 | Method and device for reconstructing visual image by utilizing electroencephalogram signal |
Non-Patent Citations (5)
| Title |
|---|
| GDTOP818: "[生成对抗网络GAN入门指南]（4）DCGAN 深度卷积生成对抗网络", 《CSDN》, pages 1 - 3 * |
| 吕立波;: "基于电子对抗原理的计算机防辐射泄密措施", 计算机安全, no. 03 * |
| 秦剑等: ""基于生成对抗网络的信号重构"", 《万方学位论文》, pages 1 - 80 * |
| 邓方等: "《智能计算与信息处理》", 30 June 2020, 北京理工大学出版社, pages: 44 - 45 * |
| 高成思;陈维伟;王颖;: "一种针对多核神经网络处理器的窃取攻击", 信息安全学报, no. 03 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108680796A (en) * | 2018-05-17 | 2018-10-19 | 集美大学 | Electromagnetic information leakage detecting system and method for computer display |
| CN108680796B (en) * | 2018-05-17 | 2023-09-19 | 集美大学 | Electromagnetic information leakage detection system and method for computer display |
| CN112818876A (en) * | 2021-02-04 | 2021-05-18 | 成都理工大学 | Electromagnetic signal extraction and processing method based on deep convolutional neural network |
| CN112818876B (en) * | 2021-02-04 | 2022-09-20 | 成都理工大学 | Electromagnetic signal extraction and processing method based on deep convolutional neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109086700B (en) | Radar one-dimensional range profile target identification method based on deep convolutional neural network | |
| Alnujaim et al. | Generative adversarial networks for classification of micro-Doppler signatures of human activity | |
| Yonel et al. | Deep learning for passive synthetic aperture radar | |
| CN110361778B (en) | Seismic data reconstruction method based on generation countermeasure network | |
| CN109522857B (en) | People number estimation method based on generation type confrontation network model | |
| CN106295694B (en) | Face recognition method for iterative re-constrained group sparse representation classification | |
| CN112966667B (en) | Method for identifying one-dimensional distance image noise reduction convolution neural network of sea surface target | |
| CN112818764B (en) | Low-resolution image facial expression recognition method based on feature reconstruction model | |
| CN108198147A (en) | A kind of method based on the multi-source image fusion denoising for differentiating dictionary learning | |
| CN105844279A (en) | Depth learning and SIFT feature-based SAR image change detection method | |
| CN109859285A (en) | Electrical impedance images method for reconstructing based on empty convolutional network | |
| CN108460391A (en) | Based on the unsupervised feature extracting method of high spectrum image for generating confrontation network | |
| CN109711401A (en) | A kind of Method for text detection in natural scene image based on Faster Rcnn | |
| CN114781435A (en) | Power electronic circuit fault diagnosis method based on improved Harris eagle optimization algorithm optimized variation modal decomposition | |
| CN112232129A (en) | Electromagnetic information leakage signal simulation system and method based on generation countermeasure network | |
| CN111414928A (en) | Method, device and equipment for generating face image data | |
| CN113627597A (en) | Countermeasure sample generation method and system based on general disturbance | |
| CN111652246B (en) | Image self-adaptive sparsization representation method and device based on deep learning | |
| CN117271979A (en) | Deep learning-based equatorial Indian ocean surface ocean current velocity prediction method | |
| CN117036901A (en) | Small sample fine adjustment method based on visual self-attention model | |
| Vetter et al. | Sourcerer: Sample-based maximum entropy source distribution estimation | |
| CN115482434A (en) | Small sample high-quality generation method based on multi-scale generation countermeasure network | |
| CN111931412A (en) | Underwater target noise LOFAR spectrogram simulation method based on generative countermeasure network | |
| CN115879513B (en) | Hierarchical standardization method and device for data and electronic equipment | |
| CN113807421B (en) | Feature map processing method of attention module based on pulse sending cortex model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |