CN116522242A - Radiation source signal open set identification method based on diffusion model - Google Patents
Radiation source signal open set identification method based on diffusion model Download PDFInfo
- Publication number
- CN116522242A CN116522242A CN202310486489.7A CN202310486489A CN116522242A CN 116522242 A CN116522242 A CN 116522242A CN 202310486489 A CN202310486489 A CN 202310486489A CN 116522242 A CN116522242 A CN 116522242A
- Authority
- CN
- China
- Prior art keywords
- radiation source
- diffusion
- distribution
- data
- diffusion model
- 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.)
- Granted
Links
- 238000009792 diffusion process Methods 0.000 title claims abstract description 64
- 238000000034 method Methods 0.000 title claims abstract description 46
- 230000005855 radiation Effects 0.000 title claims abstract description 41
- 238000001514 detection method Methods 0.000 claims abstract description 14
- 230000002441 reversible effect Effects 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 230000001186 cumulative effect Effects 0.000 claims description 3
- 238000005315 distribution function Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000012549 training Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 6
- 238000012545 processing Methods 0.000 abstract description 4
- 230000006870 function Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/10—Pre-processing; Data cleansing
-
- 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
- G06N3/047—Probabilistic or stochastic networks
-
- 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
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Probability & Statistics with Applications (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
A radiation source signal identification method based on a diffusion model relates to a radiation source signal open set identification method. The invention aims to solve the problems of poor detection effect outside the distribution of the traditional open set recognition algorithm, high algorithm complexity and poor recognition function robustness. The method is characterized in that a diffusion model is adopted to reconstruct the radiation source signals, a threshold value is set according to the statistical distribution characteristics of the reconstruction of the known class data set to distinguish the data outside the distribution, and then a low-complexity classifier is used to obtain a recognition result, so that the accuracy and the robustness of the radiation source signal recognition under the open-set scene are effectively improved. The invention belongs to the technical field of digital signal processing.
Description
Technical Field
The invention relates to a radiation source signal open set identification method, and belongs to the technical field of digital signal processing.
Background
Specific radiation source individual open set identification (SEOID), a technique aimed at identifying unknown sources of radiation, is widely used in a number of fields where SEOID can help the military identify the identity of an enemy source of radiation for effective interference and impact. In the field of communication security, the SEOID may be used to detect an illegal transmission source, so as to ensure the security of a communication system. In the field of unmanned aerial vehicle supervision, the SEOID can identify an emission source of the unmanned aerial vehicle so as to ensure the legality and safety of the unmanned aerial vehicle. In the field of Internet of things security, the SEOID can be used for protecting the security of Internet of things equipment, identifying a malicious emission source and taking corresponding measures. The traditional identification method is mainly based on the known radiation source signal model for classification, and is difficult to process the identification of unknown radiation source signals. However, the application of SEOID is also currently challenged, for example, in complex electromagnetic environments, radiation source data with low signal-to-noise ratio may have a serious negative impact on the identification capability of SEOID. It is therefore necessary to develop more robust algorithms to improve the accuracy and reliability of the SEOID.
The diffusion model is an efficient generation model that uses a reversible neural network (diffusion network) to model the diffusion process, takes initial data points as input, then updates the data points at each time step, and calculates the probability density function of the updated data points, the basic idea being to consider the data as superimposed by a number of diffusion processes, each corresponding to a dimension of the data. Diffusion models can process high-dimensional data and be used to generate high-quality images, sounds, etc., and have found wide application in many fields, such as computer vision, audio processing, natural language processing, etc. The diffusion model is applied to the open set identification field, and a good effect is achieved.
Disclosure of Invention
The invention aims to solve the problems of poor detection effect outside distribution, high algorithm complexity and poor robustness of a recognition function of the traditional open set recognition algorithm, and further provides a radiation source signal recognition method based on a diffusion model.
The technical scheme adopted by the invention for solving the problems is as follows: the method comprises the following steps:
preprocessing known radiation source data to obtain a data set formed by IQ signals;
inputting the data set into a network model of the diffusion model for training to obtain a trained diffusion model and classifier;
step three, setting a threshold value of the detection module outside the distribution according to the reconstruction error distribution of the known data;
step four, the received radiation source signals are preprocessed and then input into a diffusion module, and whether the radiation source signals belong to a known class is judged by a distribution external detection module;
and fifthly, directly outputting a result belonging to an unknown class, and recognizing the output of the inverse diffusion module as the input of a classifier to obtain a classification result when the result belongs to a known class.
Further, in the second step, the diffusion module gradually points to the data point x through T iterations 0 Adding Gaussian noise o t N (0, 1), due to data x in the forward process t Only from the previous time x t-1 In relation, this process can be regarded as a markov process as follows:
in the formulas (1) and (2), α t Method for representing noise added in the t-th diffusion step, I being the identity matrix, x during forward diffusion t Is iteratively generated by the following formula:
beta in formula (3) t ＝1-α t ，β t Is an unclonable parameter, is set at [0.0001,0.02 ]]Time step t;
the reverse diffusion module is used for recovering the original data, and the reverse Markov distribution q (x) cannot be directly obtained t-1 |x t ) Introduction of neural network p θ Fitting the distribution:
p θ (x t-1 |x t )＝N(x t-1 ；μ θ (x t ,t),∑ θ (x t ,t)) (5)，
in equations (4) and (5), Σ θ (x t T) is set to a constant beta t Mean mu θ (x t T) is a target of network fitting;
in the back diffusion process, x is used as t Predicting t-1 step dataThe method comprises the following steps:
in the formula (6) of the present invention,`o θ (x t t) represents the noise of the network prediction.
Further, in the third step, the threshold setting is performed according to the reconstruction error distribution of the known class data set, and the following method based on a normal distribution cumulative distribution function is adopted:
firstly, calculating the mean mu and the variance sigma of the reconstruction errors of the known class data through a diffusion model, and setting an outlier proportion p, wherein the outlier proportion p is set to be 5%;
and then find the Threshold value Threshold by the inverse function of the data distribution:
Threshold＝μ+σ×erf -1 (1-p0。
further, in the fourth step, the received radiation source signal is preprocessed and then input into a trained diffusion model to obtain a reconstructed signal, the reconstructed signal and the received signal are input into an out-of-distribution detection module, a reconstruction error is calculated, and whether the reconstructed signal belongs to a known class or not is judged through the threshold value set in the third step.
Further, in the fifth step, if the input signal is determined to be of an unknown class, the recognition result is directly output; if the input signal is judged to be of a known class, the reconstructed signal is sent to a classifier for known class identification, and a classification result is output.
The beneficial effects of the invention are as follows:
1. according to the invention, whether one sample belongs to a known class is evaluated by using the distribution detection module based on the diffusion model reconstruction error, so that the unknown class problem existing in the practical application can be better dealt with;
2. the invention can effectively process noise and deformation in an input signal by using the diffusion module and the inverse diffusion module, thereby having better robustness when processing noisy data;
3. the method is characterized in that a diffusion model is adopted to reconstruct the radiation source signals, a threshold value is set according to the statistical distribution characteristics of the reconstruction of the known class data set to distinguish the data outside the distribution, and then a low-complexity classifier is used to obtain a recognition result, so that the accuracy and the robustness of the radiation source signal recognition under the open-set scene are effectively improved.
Drawings
FIG. 1 is a schematic flow chart of the method of the present invention.
FIG. 2 is a schematic view of the overall structure of the present invention;
FIG. 3 is a graph of threshold discrimination effects of a distribution detection module;
FIG. 4 is a schematic diagram of a network architecture of a classifier;
FIG. 5 is a schematic diagram showing the comparison of the recognition effect of the present invention with other methods under different signal to noise ratios;
FIG. 6 is a schematic diagram showing the comparison of the recognition effect of the present invention with other methods under different opening degrees.
Detailed Description
The first embodiment is as follows: the steps of a radiation source signal identification method based on a diffusion model according to the present embodiment are as follows:
preprocessing known radiation source data to obtain a data set formed by IQ signals;
inputting the data set into a network model of the diffusion model for training to obtain a trained diffusion model and classifier;
step three, setting a threshold value of the detection module outside the distribution according to the reconstruction error distribution of the known data;
step four, the received radiation source signals are preprocessed and then input into a diffusion module, and whether the radiation source signals belong to a known class is judged by a distribution external detection module;
and fifthly, directly outputting a result belonging to an unknown class, and recognizing the output of the inverse diffusion module as the input of a classifier to obtain a classification result when the result belongs to a known class.
The data preprocessing portion in this embodiment mainly includes normalizing the amplitude and phase of complex data, and the complex form of the signal sample can be represented by IQ components as:
S i ＝Re i +jIm i
here Re i Representing the real part of the ith sample point, im i Representing the imaginary part of the i-th sample point.
The data preprocessing part mainly comprises normalization of amplitude and phase of complex data, and complex form of signal samples can be expressed as IQ components:
the second embodiment is as follows: description is made with reference to fig. 1 to 6In the second step of the radiation source signal identification method based on the diffusion model according to the present embodiment, the diffusion model gradually points to the data point x through T iterations 0 Adding Gaussian noise o t N (0, 1), due to data x in the forward process t Only from the previous time x t-1 In relation, this process can be regarded as a markov process as follows:
in the formulas (1) and (2), α t Method for representing noise added in the t-th diffusion step, I being the identity matrix, x during forward diffusion t Is iteratively generated by the following formula:
beta in formula (3) t ＝1-α t ，β t Is an unclonable parameter, is set at [0.0001,0.02 ]]Time step t;
the reverse diffusion module is used for recovering the original data, and the reverse Markov distribution q (x) cannot be directly obtained t-1 |x t ) Introduction of neural network p θ Fitting the distribution:
p θ (x t-1 |x t )＝N(x t-1 ；μ θ (x t ，t)，∑ θ (x t ，t)) (5)，
in equations (4) and (5), Σ θ (x t T) is set to a constant beta t Mean mu θ (x t T) is a target of network fitting;
in the back diffusion process, x is used as t Predicting t-1 step dataThe method comprises the following steps:
in the formula (6) of the present invention,`o θ (x t t) represents the noise of the network prediction.
And a third specific embodiment: in the third step of the radiation source signal identifying method based on the diffusion model according to the present embodiment, the threshold setting according to the reconstruction error distribution of the known class data set is performed by using the following method based on the normal distribution cumulative distribution function:
firstly, calculating the mean mu and the variance sigma of the reconstruction errors of the known class data through a diffusion model, and setting an outlier proportion p, wherein the outlier proportion p is set to be 5%;
and then find the Threshold value Threshold by the inverse function of the data distribution:
Threshold＝μ+σ×erf -1 (1-p)。
the specific embodiment IV is as follows: referring to fig. 1 to 6, a description is given of the present embodiment, in which in step four of the radiation source signal identifying method based on the diffusion model, the received radiation source signal is obtained, preprocessed, and then input into the trained diffusion model, a reconstructed signal is obtained, the reconstructed signal and the received signal are input into the out-of-distribution detection module, the reconstruction error is calculated, and whether the reconstructed signal belongs to the known class is determined through the threshold set in step three.
Fifth embodiment: in the fifth step of the radiation source signal identifying method based on the diffusion model according to the present embodiment, if the input signal is determined to be of an unknown type, the identifying result is directly output, as described with reference to fig. 1 to 6; if the input signal is judged to be of a known class, the reconstructed signal is sent to a classifier for known class identification, and a classification result is output.
Principle of operation
The invention inputs the signal of the additional radiation source after data preprocessing into the diffusion module to obtain the signal after noise adding, and transmits the signal to the inverse diffusion module to obtain the reconstruction signal, wherein the network parameter of the inverse diffusion module is provided by the trained U-net, the reconstruction signal and the input signal calculate the reconstruction error in the distribution detection module and judge whether the reconstruction error is the known class according to the threshold value, if the reconstruction error is the known class, the output of the inverse diffusion module is used as the input of the classifier, thus obtaining the classification result.
The present invention is not limited to the preferred embodiments, but is capable of modification and variation in detail, and other embodiments, such as those described above, of making various modifications and equivalents will fall within the spirit and scope of the present invention.
Claims (5)
1. A radiation source signal identification method based on a diffusion model is characterized by comprising the following steps of: the radiation source signal identification method based on the diffusion model comprises the following steps:
preprocessing known radiation source data to obtain a data set formed by IQ signals;
inputting the data set into a network model of the diffusion model for training to obtain a trained diffusion model and classifier;
step three, setting a threshold value of the detection module outside the distribution according to the reconstruction error distribution of the known data;
step four, the received radiation source signals are preprocessed and then input into a diffusion module, and whether the radiation source signals belong to a known class is judged by a distribution external detection module;
and fifthly, directly outputting a result belonging to an unknown class, and recognizing the output of the inverse diffusion module as the input of a classifier to obtain a classification result when the result belongs to a known class.
2. A radiation source signal identification method based on a diffusion model according to claim 1, characterized in that: gradually introducing the data point x into the diffusion module through T iterations in the second step 0 Adding Gaussian noise o t N (0, 1), due to data x in the forward process t Only from the previous time x t-1 In relation, this process can be regarded as a markov process as follows:
in the formulas (1) and (2), α t Method for representing noise added in the t-th diffusion step, I being the identity matrix, x during forward diffusion t Is iteratively generated by the following formula:
beta in formula (3) t ＝1-α t ，β t Is an unclonable parameter, is set at [0.0001,0.02 ]]Time step t;
the reverse diffusion module is used for recovering the original data, and the reverse Markov distribution q (x) cannot be directly obtained t-1 |x t ) Introduction of neural network p θ Fitting the distribution:
p θ (x t-1 |x t )＝N(x t-1 ；μ θ (x t ,t),∑ θ (x t ,t)) (5)，
in equations (4) and (5), Σ θ (x t T) is set to a constant beta t Mean mu θ (x t T) is a target of network fitting;
in the back diffusion process, x is used as t Predicting t-1 step dataThe method comprises the following steps:
in the formula (6) of the present invention,`o θ (x t t) represents the noise of the network prediction.
3. A radiation source signal identification method based on a diffusion model according to claim 1, characterized in that: in the third step, the threshold value setting is carried out according to the reconstruction error distribution of the known class data set, and the following method based on a normal distribution cumulative distribution function is adopted:
firstly, calculating the mean mu and the variance sigma of the reconstruction errors of the known class data through a diffusion model, and setting an outlier proportion p, wherein the outlier proportion p is set to be 5%;
and then find the Threshold value Threshold by the inverse function of the data distribution:
Threshold＝μ+σ×erf -1 (1-p)。
4. a radiation source signal identification method based on a diffusion model according to claim 1, characterized in that: and step four, acquiring a diffusion model after pretreatment of the received radiation source signals and inputting the radiation source signals into a trained diffusion model to obtain a reconstruction signal, inputting the reconstruction signal and the received signals into an out-of-distribution detection module, calculating a reconstruction error, and judging whether the reconstruction error belongs to a known class or not through the threshold value set in the step three.
5. A radiation source signal identification method based on a diffusion model according to claim 1, characterized in that: step five, if the input signal is judged to be of unknown type, directly outputting the identification result; if the input signal is judged to be of a known class, the reconstructed signal is sent to a classifier for known class identification, and a classification result is output.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310486489.7A CN116522242B (en) | 2023-04-28 | 2023-04-28 | Radiation source signal open set identification method based on diffusion model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310486489.7A CN116522242B (en) | 2023-04-28 | 2023-04-28 | Radiation source signal open set identification method based on diffusion model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116522242A true CN116522242A (en) | 2023-08-01 |
| CN116522242B CN116522242B (en) | 2024-01-26 |
Family
ID=87404240
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310486489.7A Active CN116522242B (en) | 2023-04-28 | 2023-04-28 | Radiation source signal open set identification method based on diffusion model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116522242B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113705446A (en) * | 2021-08-27 | 2021-11-26 | 电子科技大学 | Open set identification method for individual radiation source |
| CN114580534A (en) * | 2022-03-03 | 2022-06-03 | 武汉理工大学 | Industrial data anomaly detection method and device, electronic equipment and storage medium |
| CN114970638A (en) * | 2022-06-10 | 2022-08-30 | 中国人民解放军战略支援部队信息工程大学 | Radar radiation source individual open set identification method and system |
| CN115169415A (en) * | 2022-07-25 | 2022-10-11 | 中国人民解放军海军航空大学 | Communication radiation source open set identification method and system |
| CN115331055A (en) * | 2022-08-15 | 2022-11-11 | 天津大学 | Image open set identification method based on comparison thought and integrated discrimination |
-
2023
- 2023-04-28 CN CN202310486489.7A patent/CN116522242B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113705446A (en) * | 2021-08-27 | 2021-11-26 | 电子科技大学 | Open set identification method for individual radiation source |
| CN114580534A (en) * | 2022-03-03 | 2022-06-03 | 武汉理工大学 | Industrial data anomaly detection method and device, electronic equipment and storage medium |
| CN114970638A (en) * | 2022-06-10 | 2022-08-30 | 中国人民解放军战略支援部队信息工程大学 | Radar radiation source individual open set identification method and system |
| CN115169415A (en) * | 2022-07-25 | 2022-10-11 | 中国人民解放军海军航空大学 | Communication radiation source open set identification method and system |
| CN115331055A (en) * | 2022-08-15 | 2022-11-11 | 天津大学 | Image open set identification method based on comparison thought and integrated discrimination |
Non-Patent Citations (1)
| Title |
|---|
| JONATHAN HO 等: "Denoising Diffusion Probabilistic Models", NEURIPS 2020, pages 1 - 12 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116522242B (en) | 2024-01-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Lerch-Hostalot et al. | Unsupervised steganalysis based on artificial training sets | |
| CN108304877B (en) | Physical layer channel authentication method based on machine learning | |
| CN110348475B (en) | Confrontation sample enhancement method and model based on spatial transformation | |
| CN112087774B (en) | Communication radiation source individual identification method based on residual error neural network | |
| CN113014524B (en) | Digital signal modulation identification method based on deep learning | |
| CN113205140B (en) | Semi-supervised specific radiation source individual identification method based on generative countermeasure network | |
| Hong-hai et al. | Radar emitter multi-label recognition based on residual network | |
| Qi et al. | Detection tolerant black-box adversarial attack against automatic modulation classification with deep learning | |
| CN113962381A (en) | Weak signal data enhancement method | |
| Zhao et al. | Radar HRRP target recognition via semi-supervised multi-task deep network | |
| CN111680737A (en) | Radar radiation source individual identification method under differential signal-to-noise ratio condition | |
| Ghasemzadeh et al. | GGCNN: An efficiency-maximizing gated graph convolutional neural network architecture for automatic modulation identification | |
| CN117809168B (en) | Method and device for detecting inherent attribute characteristics based on underwater target | |
| CN112990357B (en) | Black box video countermeasure sample generation method based on sparse disturbance | |
| CN114943246A (en) | Semi-supervised specific radiation source identification method based on bispectrum feature extraction | |
| CN113095218B (en) | Hyperspectral image target detection algorithm | |
| CN116522242B (en) | Radiation source signal open set identification method based on diffusion model | |
| CN114024808A (en) | Modulation signal identification method and system based on deep learning | |
| CN116644439A (en) | Model safety assessment method based on denoising diffusion model | |
| CN112213697A (en) | Feature fusion method for radar deception jamming recognition based on Bayesian decision theory | |
| CN104463222B (en) | Polarimetric SAR image classification method based on feature vector distribution characteristics | |
| Wu et al. | High-confidence sample augmentation based on label-guided denoising diffusion probabilistic model for active deception jamming recognition | |
| CN112069987B (en) | Interference type automatic identification method based on statistical manifold optimization dimension reduction | |
| CN116055026A (en) | ARMA model-based particle filter side channel attack noise reduction preprocessing method | |
| CN115546003A (en) | Back door watermark image data set generation method based on confrontation training network |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |