CN112215054B - Depth generation countermeasure method for denoising underwater sound signal - Google Patents
Depth generation countermeasure method for denoising underwater sound signal Download PDFInfo
- Publication number
- CN112215054B CN112215054B CN202010733300.6A CN202010733300A CN112215054B CN 112215054 B CN112215054 B CN 112215054B CN 202010733300 A CN202010733300 A CN 202010733300A CN 112215054 B CN112215054 B CN 112215054B
- Authority
- CN
- China
- Prior art keywords
- model
- sample
- training
- countermeasure
- data
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 30
- 230000005236 sound signal Effects 0.000 title claims abstract description 15
- 238000000605 extraction Methods 0.000 claims abstract description 7
- 238000005070 sampling Methods 0.000 claims abstract description 4
- 238000013507 mapping Methods 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 11
- 230000008485 antagonism Effects 0.000 claims description 4
- 239000011159 matrix material Substances 0.000 claims description 2
- 238000005192 partition Methods 0.000 claims description 2
- 230000000694 effects Effects 0.000 description 7
- 238000013135 deep learning Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 238000000354 decomposition reaction Methods 0.000 description 5
- 238000005457 optimization Methods 0.000 description 2
- 238000001228 spectrum Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000739 chaotic effect Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000011946 reduction process 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
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
-
- 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
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Signal Processing (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Complex Calculations (AREA)
Abstract
The invention discloses a deep generation countermeasure method for denoising underwater sound signals, and belongs to the technical field of underwater sound signal denoising. Firstly, sampling and feature extraction are carried out on an original underwater sound signal, then the extracted signal is sent into a Gauss-limited Boltzmann machine, and semi-supervised pre-training of a probability generation model is carried out; and finally, constructing a deep generation countermeasure model, sending the data generated in the probability generation model and the real label data stream into a Bernoulli limited Boltzmann machine countermeasure model, and performing supervised training. Aiming at the characteristic extraction characteristic of the underwater sound signal, the generation countermeasure model is introduced into the probability model of the limited Boltzmann machine, so that the problems of strong dependence and overfitting of the limited Boltzmann machine in the training process caused by the complex signal carried by the underwater sound are effectively eliminated, and the training model has stronger self-applicability.
Description
Technical Field
The invention belongs to the technical field of underwater sound signal noise reduction, and can effectively reproduce an original useful signal from an underwater sound signal with a large signal-to-noise ratio.
Background
In the existing underwater sound signal denoising, a traditional denoising method, a time domain-based modal decomposition method and a frequency domain-based overall modal decomposition method need to set some empirical parameters in advance, so that the denoising process depends on empirical values, the classical modal decomposition method can complete the denoising process without setting functions in advance, and mode mixing and boundary effects are easy to generate in the decomposition process. In order to overcome boundary mixing, the underwater acoustic signal denoising method based on CEEMDAN, refined composite multi-scale dispersion entropy and wavelet threshold denoising has a good effect on the aspect of analyzing the complexity of chaotic signals, but has no good robustness.
In recent years, a deep learning algorithm is introduced in the underwater sound denoising field to increase the robustness of the system, a self-coding network is a typical deep learning algorithm, independence between signals and noise is not assumed in the denoising algorithm, so that the robustness of the denoising system is superior to that of a traditional denoising method, a boltzmann machine principle is added in the self-coding network to further enhance the robustness of the denoising self-coding network, and a network model is more robust. A new breakthrough in the field of deep learning generation modeling is generation of a antagonism network, which is successful in the field of computer vision, can generate vivid images, and has good effects on improving model parameters and optimizing algorithms on the basis of generation of antagonism. The generation of the confrontation model utilizes the mutual optimization effect of the generation model and the confrontation model, can obtain good effect under a limited number of training sets, and effectively solves the limitation of underwater sound denoising in the aspect of deep learning application.
In the underwater acoustic signal noise reduction process, when corresponding parameters need to be set in advance, namely under the condition of high experience requirements and the robustness of the system is poor, the traditional noise reduction method is not applicable any more, so that the advantage is displayed when deep learning is used for noise reduction.
Disclosure of Invention
In order to solve the problem of small sample training specific to underwater acoustic signals, a deep generation countermeasure method is provided for an underwater acoustic signal denoising technology. The method adds sample statistical characteristics in the training of the limiting Boltzmann network on the samples, so that the training effect on small samples is good. And aiming at the characteristics of the underwater sound and the characteristics of the underwater sound background noise, the generation of the confrontation network is further optimized, so that the training model is suitable for denoising of the underwater sound channel.
The technical scheme adopted by the invention for solving the technical problem is as follows: a depth generation countermeasure method for denoising underwater acoustic signals is characterized by comprising the following steps:
Step 1: the method adopts an MFCC feature extraction method, wherein Mel cepstrum coefficients (MFCC for short) are cepstrum parameters extracted in a Mel scale frequency domain, the Mel scale describes the nonlinear characteristic of human ear frequency, and the relationship between the Mel cepstrum coefficients and the frequency can be approximately expressed by the following formula:
where f is the sampling frequency.
And 2, step: and sending the extracted signals into a Gauss-limited Boltzmann machine for semi-supervised pre-training of a probability generation model.
The probability generation model is formed by stacking Gauss limited Boltzmann machines, firstly, the depth model obtains an initial value of a depth network hyperparameter by using data without labels and carrying out unsupervised pre-training layer by layer from the bottom layer upwards by using a Gauss limited Boltzmann machine algorithm, and after the unsupervised pre-training, the network carries out supervised training by using the data with the labels to adjust a weight value. The Gaussian limited Boltzmann machine is a generating random network, which consists of a visible layer and a hidden layer, wherein the weight theta of the network consists of a connecting weight matrix omega of the visible layer and the hidden layer, a bias vector c of the visible layer and a bias vector b of the hidden layer, and when a group of visible layer state v and hidden layer state h are given, the distribution of an energy function and a likelihood function of the limited Boltzmann machine is expressed as follows:
Wherein v isi∈{0,1}；hj∈{0,1}；
Is a partition function, when one of the visible layer and the hidden layer is in a fixed state, the conditional probability distribution of the restricted Boltzmann machine can be expressed as
Wherein,
the generation part generates the distribution of x samples under the mapping of a group of prior distribution Z by using a sample Z, the generator is an effective mapping which can simulate real data distribution to generate new samples related to a training set, and the generation model learns the mapping from input data streams to output data streams instead of the traditional mapping from input to output.
And 3, step 3: constructing a deep generation countermeasure model, sending data generated in the probability generation model and a real label data stream into a Bernoulli limited Boltzmann machine countermeasure model, and performing supervised training;
the method is characterized in that the training process is layer-by-layer training, each Boltzmann machine is trained independently and then superposed, and the defects of slow updating of back propagation parameters, large errors and the like caused by complex structures between layers are effectively avoided. The probability generation model is a network with no connection in the layer and full connection in the layer, the number of the characteristic values of the input layer of the network is equal to that of the characteristic values of the output layer of the network, and four layers of Gauss limited Boltzmann networks are superposed in the middle.
The Bernoulli limited Boltzmann machine countermeasure model is an optimization part of the whole model, the mode of generating part learning optimal mapping is completed through countermeasure training, a self-coding model in the countermeasure model adopts the Bernoulli limited Boltzmann machine model, probability judgment is well added to data training and fine tuning, a discrimination layer is added finally, the discrimination layer is a binary classifier, the input data of the countermeasure model are two, one is from real sample distribution X, and the other is from real sample distribution simulated by a generator
The function of the discriminator is to determine the true sample distribution X as the true data, and the generator generates the sample distribution
For false data, the generator continuously optimizes according to the output result of the competitor until the competitor cannot judge whether the input data is real data or generated data, which shows that the distribution of the input data is very close to that of the real data, and the training becomes the maximum and minimum game training for generating the competitor model, and the purpose is shown in formula 6:
on the basis, we can add some extra condition values to perform mapping and classification, and in the denoising model, the extra condition values are noisy samples x cFormula 6 may become represented by formula 7 after addition:
for stability in the training process, the generator and the countermeasure are trained separately, and the minimum values of the two functions are optimized respectively as shown in equations 8 and 9:
the method comprises the steps that a discrimination network model is mainly responsible for antagonism, when a discrimination system easily discriminates that a clean sample is 1 at the beginning and a generated sample is 0, when the generated sample of the discrimination generator is 0, the generator starts to optimize parameters of the generator, so that a value mapped by the generator from an input sample is close to a value of a real sample, when the generated sample is close to the real sample and cannot be discriminated by the discriminator, the discriminator starts to optimize the parameters of the discriminator, so that the generated sample and the real sample can be well discriminated, a two-dimensional array is input into the generation model, and when the generated sample and a sample with noise are input, the discriminator discriminates 0; when a clean sample and a noisy sample are input, the discrimination result is 1.
The beneficial effects of the invention are: aiming at the characteristic extraction characteristic of the underwater sound signal, a generation countermeasure model is introduced into the probability model of the limited Boltzmann, so that the problems of strong dependence and overfitting of the limited Boltzmann machine in the training process caused by the complex signal carried by the underwater sound are effectively eliminated, and the training model has stronger self-applicability.
The invention provides a deep generation countermeasure algorithm for improving the signal-to-noise ratio in the underwater acoustic signal feature extraction technology, a traditional underwater acoustic signal noise reduction method, such as a wavelet denoising algorithm and an integrated empirical mode decomposition algorithm, has certain assumed conditions for samples before denoising, but certain assumed conditions cannot be completely met in the actual underwater environment, a coding model for generating the countermeasure network is a typical deep learning network model which can denoise noisy signals well without assumption independence, but a convolutional neural network is used for generating the network and the countermeasure network in a common generation countermeasure denoising model, so that the data with a large number of training samples has good robustness, but for the underwater acoustic signals of small samples, a Gauss limited Boltzmann machine and a Bernoulli first Boltzmann machine with probability statistical properties are respectively added in a generation unit and a countermeasure unit for generating the countermeasure model, firstly, optimizing network parameters by utilizing semi-supervised learning to form a deep generation confrontation network model.
The deep generation countermeasure network provided by the invention hopes to eliminate the noise of the original noisy signal to obtain a clean underwater sound signal, the main sources of the underwater sound noise are ocean background noise and propeller noise, including linear noise and nonlinear noise, and the underwater acoustic signal training samples are limited in quantity, the deep generation countermeasure method learning data is not a simple mapping relation from input to output of the learning data, but a mapping relation between the data statistical characteristics and the statistical characteristics of a given sample is learned, then, supervised learning is carried out by using a countermeasure method, the statistical characteristics of the input underwater sound signals are continuously updated, it is also possible to reconstruct the underwater sound signal well, in a mixed linear and non-linear system, and with a small number of training sets, so that it is close to a clean underwater sound signal, and the obtained deep generation countermeasure model with the data statistical characteristics has strong robustness and self-adaptive capacity.
The present invention will be described in detail with reference to the following embodiments.
Drawings
FIG. 1 is a discriminant model of a system
Fig. 2 is a diagram of generating a confrontation network model.
Fig. 3 red is a time domain diagram of the underwater sound after denoising by applying the depth generation countermeasure method, blue is a time domain diagram of a signal with noise, and green is a time domain diagram of a clean signal.
FIG. 4 shows the left graph of the spectrum of a clean underwater acoustic signal and the right graph of the spectrum of a signal denoised by a depth-generation countermeasure method
Detailed Description
Step 3, adding the data generated by the generator into the noisy data, sending the data and the original clean noisy data into a discrimination model for discrimination, starting the discriminator to discriminate well, wherein the data of the generated model is a false sample and the output is 0, the data of the original clean noisy data is a true sample and the output is 1, according to the result judged by the discriminator, the generator starts to simulate the generated data thereof to make the data as close as possible to the real data, thus, until the discriminator has no way to discriminate, the data generated by the generator is sent to the discriminator again, the discriminator can not discriminate, when the input is two samples, the output is 0.5, after a period of training, the generated sample can be well distinguished to be 0, and the steps 3 and 4 are repeated until the output value of the discriminator is stabilized at 0.5, and the model generated by the generator is a clean sample.
As can be seen from fig. 3 and 4, in the embodiment, under the same signal-to-noise ratio, for the denoising effect of the underwater acoustic signal of the small sample, due to the traditional denoising method, the noisy signal time domain diagram of the trained denoising model reaches a very high coincidence degree with the signal time domain diagram of the clean sample, and is found in a frequency domain, the energy of the clean sample is well concentrated at one point, the energy of the denoising model has a deviation, but the energy is concentrated at a specific point, and the frequency shift is found to be a fixed value according to a plurality of experiments, so that an empirical value of the frequency shift is obtained, and the noisy speech can be well denoised.
Claims (1)
1. A depth-generated countermeasure method for denoising underwater acoustic signals is characterized by comprising the following steps:
step 1: sampling and feature extracting are carried out on the original underwater sound signal; the used feature extraction method is a Mel cepstrum coefficient feature extraction method, wherein Mel cepstrum coefficients are abbreviated as MFCC, MFCC is cepstrum parameters extracted in Mel scale frequency domain, Mel scale describes nonlinear characteristics of human ear frequency, and the relationship between the Mel cepstrum coefficients and the frequency is approximately expressed by the following formula:
wherein f is the sampling frequency;
step 2: sending the extracted signals into a Gauss limited Boltzmann machine for semi-supervised pre-training of a probability generation model;
Firstly, performing unsupervised pre-training on a depth model layer by layer from the bottom layer upwards by using data without labels by using a limited Gauss boltzmann machine algorithm to obtain an initial value of a hyper-parameter of a depth network, and after unsupervised pre-training, performing supervised training on the network by using data with labels to adjust a weight; the Gauss limited Boltzmann machine is a generating random network, which consists of a visible layer and a hidden layer, wherein a weight theta of the network consists of a connection weight matrix omega of the visible layer and the hidden layer, a bias vector c of the visible layer and a bias vector b of the hidden layer; given a set of visible layer states v, and hidden layer states h, the energy function and likelihood function distribution of a constrained boltzmann machine is represented as:
wherein v isi∈{0,1}；hj∈{0,1}；
Is a partition function, when one of the visible layer and the hidden layer is in a fixed state, the conditional probability distribution of the restricted Boltzmann machine can be expressed as
Wherein,
the generation part generates the distribution of x samples by using a sample S under the mapping of a group of prior distribution S, the generator is an effective mapping which can simulate real data distribution to generate new samples related to a training set, and the generation model learns the mapping from input data flow to output data flow instead of the traditional mapping from input to output;
And 3, step 3: constructing a deep generation countermeasure model, sending data generated in the probability generation model and a real label data stream into a Bernoulli limited Boltzmann machine countermeasure model, and carrying out supervised training;
the generation countermeasure model consists of a generation model and a countermeasure model, wherein the generation model adopts a probability generation model; the countermeasure model is a Bernoulli limited Boltzmann machine countermeasure model which is an optimized part of the whole model, the mode of generating part learning optimal mapping is completed by countermeasure training, a self-coding model in the countermeasure model adopts a Bernoulli limited Boltzmann machine model, the probability judgment is well added to data training and fine tuning,finally, a discrimination layer is added, namely a binary classifier is adopted, and input data of the countermeasure model are divided into two parts, namely a real sample distribution X from the generator and a real sample distribution simulated by the generator
The function of the discriminator is to determine the true sample distribution X as true data, and the generator generates the sample distribution
For false data, the generator continuously optimizes according to the output result of the countermeasure until the countermeasure cannot judge whether the input data is real data or generated data, which shows that the distribution of the input data is very close to that of the real data, and the training becomes the maximum and minimum game training for generating the countermeasure model, and the purpose is as shown in formula 6:
In the de-noising model, the additional condition value is a noisy sample x1Formula 6 may become represented by formula 7 after addition:
training the generator and the countermeasure device separately, and respectively optimizing the minimum value of two functions, as shown in formulas 8 and 9:
the judgment network model is mainly responsible for antagonism, the judgment system easily judges that a clean sample is 1 and a generated sample is 0 at the beginning, when the generated sample of the judgment generator is 0, the generator starts to optimize self parameters to enable a value mapped by the generator from an input sample to be close to a value of a real sample, when the generated sample is close to the real sample and the judgment device cannot accurately judge whether the generated sample is the clean sample or the generated sample, the judgment device starts to optimize the self parameters to enable the generated sample and the real sample to be well judged, a two-dimensional array is input into the generation model, and when the generated sample and a noisy sample are input, the judgment device judges that the generated sample is 0; when the clean sample and the noisy sample are input, the discrimination result is 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010733300.6A CN112215054B (en) | 2020-07-27 | 2020-07-27 | Depth generation countermeasure method for denoising underwater sound signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010733300.6A CN112215054B (en) | 2020-07-27 | 2020-07-27 | Depth generation countermeasure method for denoising underwater sound signal |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112215054A CN112215054A (en) | 2021-01-12 |
| CN112215054B true CN112215054B (en) | 2022-06-28 |
Family
ID=74059288
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010733300.6A Active CN112215054B (en) | 2020-07-27 | 2020-07-27 | Depth generation countermeasure method for denoising underwater sound signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112215054B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113541835B (en) * | 2021-06-17 | 2022-05-20 | 江苏科技大学 | Time-varying underwater acoustic channel simulation method for generating countermeasure network based on conditions |
| CN113823308B (en) * | 2021-09-18 | 2023-11-28 | 东南大学 | Method for denoising voice by using single voice sample with noise |
| CN113962381B (en) * | 2021-09-30 | 2023-02-03 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Weak signal data enhancement method |
| CN114785423B (en) * | 2022-06-20 | 2022-11-01 | 中国海洋大学三亚海洋研究院 | Method for building underwater laser digital information transmission enhancement model |
| CN116403590B (en) * | 2023-06-08 | 2023-08-18 | 青岛科技大学 | Bionic signal processing method based on wavelet transformation and generation countermeasure network |
| CN118379982A (en) * | 2024-06-27 | 2024-07-23 | 武汉普惠海洋光电技术有限公司 | High-frequency array environment noise reduction method and device |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105488466A (en) * | 2015-11-26 | 2016-04-13 | 中国船舶工业系统工程研究院 | Deep neural network and underwater sound target vocal print feature extraction method |
| CN107563428A (en) * | 2017-08-25 | 2018-01-09 | 西安电子科技大学 | Classification of Polarimetric SAR Image method based on generation confrontation network |
| CN109147817A (en) * | 2018-08-29 | 2019-01-04 | 昆明理工大学 | A kind of denoising audio feature extraction algorithm based on the limited Boltzmann machine that makes a variation |
| CN109978079A (en) * | 2019-04-10 | 2019-07-05 | 东北电力大学 | A kind of data cleaning method of improved storehouse noise reduction self-encoding encoder |
| CN109993710A (en) * | 2019-03-20 | 2019-07-09 | 西北工业大学 | A kind of underwater picture denoising method based on generation confrontation network |
| CN110490816A (en) * | 2019-07-15 | 2019-11-22 | 哈尔滨工程大学 | A kind of underwater Heterogeneous Information data noise reduction |
| CN111401236A (en) * | 2020-03-16 | 2020-07-10 | 西北工业大学 | Underwater sound signal denoising method based on self-coding neural network |
| CN111428853A (en) * | 2020-01-16 | 2020-07-17 | 东华大学 | Negative sample confrontation generation method with noise learning function |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11257259B2 (en) * | 2017-08-15 | 2022-02-22 | Siemens Healthcare Gmbh | Topogram prediction from surface data in medical imaging |
| CN108765319B (en) * | 2018-05-09 | 2020-08-14 | 大连理工大学 | Image denoising method based on generation countermeasure network |
| CN110020685A (en) * | 2019-04-09 | 2019-07-16 | 山东超越数控电子股份有限公司 | A kind of preprocess method based on adaptive-filtering and limited Boltzmann machine, terminal and readable storage medium storing program for executing |
-
2020
- 2020-07-27 CN CN202010733300.6A patent/CN112215054B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105488466A (en) * | 2015-11-26 | 2016-04-13 | 中国船舶工业系统工程研究院 | Deep neural network and underwater sound target vocal print feature extraction method |
| CN107563428A (en) * | 2017-08-25 | 2018-01-09 | 西安电子科技大学 | Classification of Polarimetric SAR Image method based on generation confrontation network |
| CN109147817A (en) * | 2018-08-29 | 2019-01-04 | 昆明理工大学 | A kind of denoising audio feature extraction algorithm based on the limited Boltzmann machine that makes a variation |
| CN109993710A (en) * | 2019-03-20 | 2019-07-09 | 西北工业大学 | A kind of underwater picture denoising method based on generation confrontation network |
| CN109978079A (en) * | 2019-04-10 | 2019-07-05 | 东北电力大学 | A kind of data cleaning method of improved storehouse noise reduction self-encoding encoder |
| CN110490816A (en) * | 2019-07-15 | 2019-11-22 | 哈尔滨工程大学 | A kind of underwater Heterogeneous Information data noise reduction |
| CN111428853A (en) * | 2020-01-16 | 2020-07-17 | 东华大学 | Negative sample confrontation generation method with noise learning function |
| CN111401236A (en) * | 2020-03-16 | 2020-07-10 | 西北工业大学 | Underwater sound signal denoising method based on self-coding neural network |
Non-Patent Citations (5)
| Title |
|---|
| Boosting Noise Robustness of Acoustic Model via Deep Adversarial Training;Bin Liu 等;《2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)》;20180913;第5034-5038页 * |
| Spectra-GANs: A New Automated Denoising Method for Low-S/N Stellar Spectra;Minglei Wu 等;《IEEE Access》;20200605;第8卷;第107912-107926页 * |
| The research of underwater target recognition method based on deep learning;Yuechao Chen 等;《2017 IEEE International Conference on Signal Processing, Communications and Computing (ICSPCC)》;20180101;第1-5页 * |
| 基于Adversarial的生成式算法研究;罗金铭;《中国硕士学位论文全文数据库 信息科技专辑》;20190215;第2019年卷(第2期);I138-1093 * |
| 生成对抗网络的改进方法与超分辨率图像重建研究;李艳春;《中国博士学位论文全文数据库 信息科技辑》;20200115;第2020年卷(第1期);I138-77 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112215054A (en) | 2021-01-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112215054B (en) | Depth generation countermeasure method for denoising underwater sound signal | |
| CN110599409B (en) | Convolutional neural network image denoising method based on multi-scale convolutional groups and parallel | |
| CN111816156B (en) | Multi-to-multi voice conversion method and system based on speaker style feature modeling | |
| CN110334580A (en) | The equipment fault classification method of changeable weight combination based on integrated increment | |
| CN107358626A (en) | A kind of method that confrontation network calculations parallax is generated using condition | |
| CN111723701B (en) | Underwater target identification method | |
| CN109242090B (en) | Video description and description consistency judgment method based on GAN network | |
| CN112307714A (en) | Character style migration method based on double-stage deep network | |
| CN112464004A (en) | Multi-view depth generation image clustering method | |
| CN113850013B (en) | Ship radiation noise classification method | |
| CN114038055A (en) | Image generation method based on contrast learning and generation countermeasure network | |
| CN106203628A (en) | A kind of optimization method strengthening degree of depth learning algorithm robustness and system | |
| CN113256508A (en) | Improved wavelet transform and convolution neural network image denoising method | |
| CN110490265A (en) | A kind of image latent writing analysis method based on two-way convolution sum Fusion Features | |
| CN114861838A (en) | Intelligent classification method for pulsatile neural brains based on neuron complex dynamics | |
| CN118196231B (en) | Lifelong learning draft method based on concept segmentation | |
| CN103824285B (en) | Image segmentation method based on bat optimal fuzzy clustering | |
| CN114897884A (en) | No-reference screen content image quality evaluation method based on multi-scale edge feature fusion | |
| Ma et al. | Enhancing the security of image steganography via multiple adversarial networks and channel attention modules | |
| CN116776093A (en) | Electroencephalogram signal data amplification algorithm based on improved generation countermeasure network | |
| CN114723733B (en) | Class activation mapping method and device based on axiom explanation | |
| CN114841887B (en) | Image recovery quality evaluation method based on multi-level difference learning | |
| CN114202056A (en) | Small sample underwater target identification method based on multi-platform auditory perception feature deep transfer learning | |
| CN112434716B (en) | Underwater target data amplification method and system based on condition countermeasure neural network | |
| CN112906815A (en) | Method for predicting human face by sound based on condition generation countermeasure 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 |