CN109544656B - Compressed sensing image reconstruction method and system based on generation countermeasure network - Google Patents
Compressed sensing image reconstruction method and system based on generation countermeasure network Download PDFInfo
- Publication number
- CN109544656B CN109544656B CN201811404831.XA CN201811404831A CN109544656B CN 109544656 B CN109544656 B CN 109544656B CN 201811404831 A CN201811404831 A CN 201811404831A CN 109544656 B CN109544656 B CN 109544656B
- Authority
- CN
- China
- Prior art keywords
- image
- discriminator
- generator
- network model
- countermeasure network
- 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 21
- 238000005259 measurement Methods 0.000 claims abstract description 52
- 238000012549 training Methods 0.000 claims abstract description 41
- 238000005070 sampling Methods 0.000 claims abstract description 18
- 238000013528 artificial neural network Methods 0.000 claims abstract description 6
- 230000006870 function Effects 0.000 claims description 40
- 239000011159 matrix material Substances 0.000 claims description 14
- 230000004913 activation Effects 0.000 claims description 11
- 235000004257 Cordia myxa Nutrition 0.000 claims description 4
- 244000157795 Cordia myxa Species 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 4
- 230000008485 antagonism Effects 0.000 claims description 3
- 238000013507 mapping Methods 0.000 abstract description 2
- 238000012360 testing method Methods 0.000 description 7
- 230000000694 effects Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000003384 imaging method Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000004445 quantitative analysis Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000004451 qualitative analysis Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
-
- 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
-
- 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
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a compressed sensing image reconstruction method based on a generation countermeasure network, which comprises the following steps: s1, constructing a generated countermeasure network model based on a neural network according to a measurement vector obtained by sampling an original image and the size of a reconstructed image, and designing an objective function for optimizing parameters of the generated countermeasure network model; s2, presetting parameters for training when the countermeasure network model is generated; s3, alternately training a generator and a discriminator by adopting a back propagation algorithm according to the objective function; s4, if the generated countermeasure network model converges, the trained network can directly realize a compressed sensing task, and the model is output as a corresponding image reconstructed by the measurement vector; otherwise, the steps S2-S4 are executed in a returning mode. The invention uses the strong mapping capability of the generator to preliminarily reconstruct the original image, and uses the countermeasure training of the generator and the discriminator to ensure that the pixel distribution of the image reconstructed by the generator is closer to the original image, thereby achieving the purpose of accurately reconstructing the original image under the low sampling rate.
Description
Technical Field
The invention relates to the technical field of image information processing, in particular to a compressed sensing image reconstruction method and system based on a generated countermeasure network.
Background
Compressed sensing (Compressed Sensing, CS) is a novel signal acquisition theory, integrates the traditional sampling and compression processes, can directly acquire measurement data far lower than Nyquist sampling rate, can reduce sampling cost and storage resources, and meanwhile, the coding end of a compressed sensing model only needs to perform linear random measurement, and the complex optimization process of reconstructing signals is completed at the decoding end.
The imaging systems reconstruct images using iterative optimization algorithms on a small number of observed measurements based on compressed sensing theory. However, these reconstruction algorithms all need to perform complex iterative operation, the reconstruction time is long, and the quality of the reconstructed image is relatively poor under the lower sampling rate, which hinders the deep development and industry application of compressed sensing imaging.
Disclosure of Invention
The invention aims to: in order to overcome the defects of the prior art, the invention provides a compressed sensing image reconstruction method based on a generation countermeasure network, which can solve the problems of longer reconstruction time and poorer quality of reconstructed images.
The technical scheme is as follows: the invention relates to a compressed sensing image reconstruction method based on a generation countermeasure network, which comprises the following steps:
s1, constructing a neural network-based generation countermeasure network model according to a measurement vector obtained by image sampling and the size of a reconstructed image, and designing an objective function for optimizing parameters of the generation countermeasure network model;
s2, presetting parameters for training when the countermeasure network model is generated;
s3, alternately training a generator and a discriminator by adopting a back propagation algorithm according to the objective function;
s4, if the generated countermeasure network model converges, the trained network can directly realize a compressed sensing task, and the model is output as a corresponding image reconstructed by the measurement vector; otherwise, the steps S2-S4 are executed in a returning mode.
Preferably, in S1, the measurement vector is expressed as:
wherein y represents a measurement vector, phi is a measurement matrix, x represents a data line vector which is changed into a data matrix of an image to be sampled after being changed into vectorization, M represents the size of the measurement vector, and N is the number of pixels of the image to be sampled.
Preferably, in the step S1, the formula for generating the countermeasure network model is:
wherein G is a generator, D is a discriminator, y is a measurement vector, and x representsThe data matrix of the image to be sampled becomes a vector of data rows that becomes vectorized,reconstruction loss->Regular for total variation->To combat losses in the discriminator network.
Preferably, substituting the reconstruction loss, the total variation regularization and the counterloss correspondence formula of the discriminator network into the formula for generating the counternetwork model to obtain objective functions for optimizing parameters of the generator and the discriminator respectively;
the objective function of the generator is expressed as:
the objective function of the discriminator is:
wherein,,the generator reconstructs an image, i and j are pixel positions of the original image, p data And p y Is the statistical distribution of the original image and the random measurement vector, and γ is the loss weight of the discriminator.
Preferably, in S3, the activation function used in training the generator is Selu, and the activation function used in training the discriminator is lrinlu.
The invention also discloses a system for realizing the compressed sensing image reconstruction method based on the generation countermeasure network, which comprises the following steps:
the network model construction module is used for constructing a neural network-based generation countermeasure network model according to the measurement vector obtained by image sampling and the size of the reconstructed image, and designing an objective function for optimizing parameters of the generation countermeasure network model;
the parameter presetting module is used for presetting parameters for training the generation of the countermeasure network model;
the generator training module is used for alternately training the generator and the discriminator by adopting a back propagation algorithm according to the objective function;
the convergence judging module is used for judging whether the generated countermeasure network model converges or not, if the generated countermeasure network model converges, the trained network can directly realize a compressed sensing task, and the model is output as a corresponding original image reconstructed by the measurement vector; otherwise, returning to the parameter preset module.
Preferably, in the network model construction module,
the measurement vector is expressed as:
wherein y represents a measurement vector, phi is a measurement matrix, x represents a data row vector which is changed after the data matrix of the image to be sampled becomes vectorized, M represents the size of the measurement vector, and N is the pixel number of the image to be sampled.
Preferably, in the network model building module, the formula for generating the countermeasure network model is:
wherein G is a generator, D is a discriminator, y is a measurement vector, x represents a data line vector which becomes vectorized after the data matrix of the original image becomes vectorized,the reconstruction is lost to the process,/>regular for total variation->To combat losses in the discriminator network.
Preferably, substituting the reconstruction loss, the total variation regularization and the counterloss correspondence formula of the discriminator network into the formula for generating the counternetwork model to obtain objective functions for optimizing parameters of the generator and the discriminator respectively;
the objective function of the generator is expressed as:
the objective function of the discriminator is:
wherein,,the generator reconstructs an image, i and j are pixel positions of the original image, p data And p y Is the statistical distribution of the original image and the random measurement vector, and γ is the loss weight of the discriminator.
Preferably, in the generator training module, the activation function used in training the generator is Selu, and the activation function used in training the discriminator is lrinlu.
The beneficial effects are that: the invention establishes a compressed sensing image reconstruction method based on a generation countermeasure network by utilizing a compressed sensing theory, and the model initially reconstructs an original image by utilizing the strong mapping capability of the generator under the condition of low sampling rate of the image, and simultaneously, further enhances the capability of the generator by utilizing countermeasure training of the generator and the discriminator, so that the pixel distribution of the image reconstructed by the generator is closer to the original image, and the purpose of accurately reconstructing the original image under the low sampling rate is achieved.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
fig. 2 is a comparison result of a plurality of algorithms for a reconstructed image and an original image of a CelebA face data set according to the present invention, wherein fig. 2a is a schematic diagram of a comparison result of a reconstructed image and an original image obtained by a ReconNet algorithm, fig. 2b is a comparison result of a reconstructed image and an original image obtained by a CSGAN algorithm, fig. 2c is a schematic diagram of a comparison result of a reconstructed image and an original image obtained by a DCGAN algorithm, and fig. 2d is a schematic diagram of a comparison result of a reconstructed image and an original image obtained by a method according to the present invention;
FIG. 3 is a graph of the mean square error comparison result of the Reconnet algorithm, the CSGAN algorithm, the DCGAN algorithm, and the reconstructed image and the original image obtained by the present algorithm, which are related in the present invention;
FIG. 4 is a graph of the results of comparing the peak signal-to-noise ratio and the image structure similarity index of the reconstructed image and the original image obtained by the Reconnet algorithm, the CSGAN algorithm and the DCGAN algorithm;
FIG. 5 is a graph of time versus results for the Reconnet algorithm, the CSGAN algorithm, the DCGAN algorithm, and the present algorithm to reconstruct an image, as referred to in the present invention;
fig. 6 is a schematic diagram of a system configuration according to the present invention.
Detailed Description
As shown in fig. 1, the present invention provides a compressed sensing image reconstruction method based on generation of an countermeasure network, comprising:
Measurement vector y=Φx+ζ, y∈r M ,Φ∈R M×N ,x∈R N Y represents a measurement vector, phi is a measurement matrix, x represents a data line vector which is changed into a vectorization after a picture data matrix is changed into vectorization, M represents the size of the measurement vector, and N is the number of pixels of an original image.
And 2, constructing and generating an countermeasure network model according to the size of the measurement vector and the size of the reconstructed image, and designing an objective function for optimizing parameters of the network model.
The size of the measurement vector determines the size of the input of the generator, and the size of the reconstructed image determines the final output size of the generator. The generating an countermeasure network model is as follows:
wherein G is a generator, D is a discriminator, and the reconstruction loss isThe total variation is regular asThe countermeasures against loss of the discriminator network are +.>Substituting the reconstruction loss, the total variation regularization and the counterloss of the discriminator network into the network model to obtain objective functions respectively optimizing parameters of the generator and the discriminator, wherein the objective functions of the generator are respectively +.>The objective function of the discriminator is The generator reconstructs an image, i and j are pixel positions of the original image, p data And p y Is the statistical distribution of the original image and the random measurement vector, and γ is the loss weight of the discriminator.
The preset parameters comprise model learning rate alpha during training, iteration times L, training batch size S, loss weight gamma of the discriminator during training of the generator, depth and layer number of the generator and the discriminator network, activation function of the generator and class of the generator of the discriminator.
And 4, alternately training the generator and the discriminator by adopting a back propagation algorithm according to the objective function.
Specifically, step 4 specifically includes:
step 41, selecting S training images { x } in training set (1) ,…,x (s) And get the corresponding measurement vector { y } (1) ,…,y (s) };
Step 42, updating the parameter ω in the discriminator by a back propagation algorithm.
Second, the generator parameter θ is fixed, and the discriminator parameter ω++ω·rmsprop (ω, d) is updated ω ). Wherein RMSProp is one of the gradient descent algorithms.
Step 43, constructing the generator loss function required by training the generator network according to the modelCalculating a loss function of S training images in a batch, fixing a discriminator parameter omega, and updating a generator parameter theta by adopting an RMSProp gradient descent algorithm: theta-alpha-RMSProp (theta, g) θ )。
Step 44, sequentially performing step 41, step 42 and step 43 on all images of the whole training set, and performing a total of L iterations.
And step 5, if the network model is converged, the trained network can directly realize the compressed sensing task of the segment opposite end, the output of the model is the corresponding image reconstructed by the measurement vector, otherwise, the step 3 is executed in a return mode, and the steps 3-5 are circulated.
Specifically, the step 5 specifically includes:
step 51, it is determined whether the generator network and the discriminator network converge. Iterative training in a networkIn the course, when the discriminator loses d ω Sum generator g θ And when the values are reduced and respectively gradually increased to a certain value, the network is judged to be converged.
Step 52, after the measurement vector obtained after sampling the measurement matrix Φ is input to the converged generator network, the output of the generator is the corresponding reconstructed image.
Step 53, if the iterative training does not converge, returning to execute step 3.
In order to verify the effect of the invention, the invention is subjected to simulation experiments, the specification of a test image is 64 multiplied by 64, a training and testing model is trained on a CelebA face data set, and related parameters are set: the model learning rate α=0.002, the iteration number l=30, the number of pictures per input s=16, γ=0.01, and the process of the network of generator and discriminator from input to output is shown in tables 1 and 2 below.
Training depth and activation function of table 1 generator
Table 2 training depth and activation function of discriminator
The evaluation of the experiment uses both qualitative and quantitative analysis methods.
Fig. 2 shows a comparison of the image reconstruction effect of the three algorithms of the present invention and ReconNet, CSGAN, DCGAN at sample block numbers of 20, 100 and 500, respectively. The sample block is the size of the measurement vector y, and as can be seen from fig. 2, the reconstruction effect of the invention is obviously better than that of the three algorithms ReconNet, CSGAN, DCGAN for the same four face images.
Regarding quantitative analysis comparisons, MSE, PSNR, and SSIM are used to evaluate image quality, and the reconstruction time (in milliseconds) of a single image is used to evaluate the reconstruction speed of the algorithm. Where MSE is the mean square error (Mean Squared Error), i.e. the average error of a single pixel in the image, PSNR is the peak signal to noise ratio (Peak Signal to Noise Ratio), SSIM is the structural similarity (structural similarity index), which is calculated as follows:
wherein range represents the dynamic range of the image pixel values, μ is the mean, σ is the variance, c 1 =(k 1 L) 2 ,c 1 =(k 1 L) 2 Is a constant for maintaining stability, L is the dynamic range of pixel values, k 1 =0.01,k 2 =0.03。
When quantitative comparison is carried out, 4 test pictures are selected from the test set, sampling is carried out by using a measurement matrix phi, the test pictures are input into each model, a reconstructed image is calculated and output through the model, the reconstructed image is compared with a corresponding original image, and corresponding MSE, PSNR and SSIM values are calculated.
Fig. 3 shows the MSE (under log2 scaling) comparison values for the algorithms of the invention and ReconNet, CSGAN, DCGAN, TWLST, LASSO, respectively, in the test set of CelebA, with an abscissa of Number of Measurements (number of samples) and an ordinate of the error per pixel, i.e. the mean square error (Reconstruction error per pixel); fig. 4 shows the PSNR and SSIM comparison values for the algorithm of the present invention and ReconNet, CSGAN, DCGAN at different sample blocks. Fig. 5 shows a comparison of the mean time (in milliseconds) for the present algorithm and ReconNet, CSGAN, DCGAN, twIST and LASSO algorithms to reconstruct an image in the mnist, fmnist and CelebA test sets, respectively.
The present invention also provides a compressed perceived image reconstruction system based on a generation countermeasure network, as shown in fig. 6, comprising:
the network model construction module is used for constructing a neural network-based generation countermeasure network model according to the measurement vector obtained by image sampling and the size of the reconstructed image, and designing an objective function for optimizing parameters of the generation countermeasure network model;
the parameter presetting module is used for presetting parameters for training the generation of the countermeasure network model;
the generator training module is used for alternately training the generator and the discriminator by adopting a back propagation algorithm according to the objective function;
the convergence judging module is used for judging whether the generated countermeasure network model converges or not, if the generated countermeasure network model converges, the trained network can directly realize a compressed sensing task, and the model is output as a corresponding original image reconstructed by the measurement vector; otherwise, returning to the parameter presetting module.
The system is realized based on a compressed sensing image reconstruction method for generating an countermeasure network, and the specific technology is similar to the method, and the invention is not repeated here. The foregoing is only a preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions easily contemplated by those skilled in the art within the scope of the present invention should be covered by the present invention
In summary, compared with the conventional compressed sensing, the method has the advantages of long reconstruction iteration time and poor reconstruction quality at a low sampling rate. The invention utilizes deep learning to establish a compressed sensing image reconstruction model based on generation of an antagonism network, enables a generator G to reconstruct an image, uses a discriminator D to improve pixel distribution of the reconstructed image, reduces reconstruction errors, enables reconstruction of a complex image at a low sampling rate to have a better effect, and has a better reconstruction effect on a more complex color image, and has certain advantages from the aspects of reconstruction errors, similarity and efficiency and visual effect.
Claims (4)
1. A compressed perceived image reconstruction method based on a generation countermeasure network, comprising:
s1, constructing a generated countermeasure network model based on a neural network according to a measurement vector obtained by sampling an original image and the size of a reconstructed image, and designing an objective function for optimizing parameters of the generated countermeasure network model;
the measurement vector y=Φx+ζ, y ε R M ,Φ∈R M×N ,x∈R N ;
Wherein y represents a measurement vector, phi is a measurement matrix, x represents a data line vector which is changed into a vectorized data matrix of an image to be sampled, M represents the size of the measurement vector, and N is the number of pixels of the image to be sampled;
the generating an countermeasure network model is as follows:
J(D,G)=min G max D [l rec (G(y))+l reg (G(y))+l GAN (G(y),D)]
wherein G is a generator, D is a discriminator, and the reconstruction loss isl reg For total variation regularization, the countermeasures of the discriminator network are +.>Substituting the reconstruction loss, the total variation regularization and the antagonism loss of the discriminator network into the network model to obtain objective functions respectively optimizing parameters of the generator and the discriminator, wherein the objective functions respectively comprise:
the objective function of the discriminator is The generator reconstructs an image, i and j are pixel positions of the original image, p data And p y Statistical distribution of original image and random measurement vector, gamma is loss weight of discriminator;
s2, presetting parameters for training when the countermeasure network model is generated;
s3, alternately training a generator and a discriminator by adopting a back propagation algorithm according to the objective function;
s4, if the generated countermeasure network model converges, the trained network can directly realize a compressed sensing task, and the model is output as a corresponding image reconstructed by the measurement vector; otherwise, the steps S2-S4 are executed in a returning mode.
2. The compressed sensing image reconstruction method based on a generation countermeasure network according to claim 1, wherein in S3, an activation function used in training the generator is Selu, and an activation function used in training the discriminator is lrinlu.
3. A system implemented in accordance with the compressed perceived image reconstruction method of claim 1 or 2, comprising:
the network model construction module is used for constructing a generated countermeasure network model based on a neural network according to the measurement vector obtained by sampling the original image and the size of the reconstructed image, and designing an objective function for optimizing parameters of the generated countermeasure network model;
the parameter presetting module is used for presetting parameters for training the generation of the countermeasure network model;
the generator training module is used for alternately training the generator and the discriminator by adopting a back propagation algorithm according to the objective function;
the convergence judging module is used for judging whether the generated countermeasure network model converges or not, if the generated countermeasure network model converges, the trained network can directly realize a compressed sensing task, and the model is output as a corresponding original image reconstructed by the measurement vector; otherwise, returning to the parameter preset module;
the measurement vector y=Φx+ζ, y ε R M ,Φ∈R M×N ,x∈R N ,
Wherein y represents a measurement vector, phi is a measurement matrix, x represents a data line vector which is changed into a vectorized data matrix of an image to be sampled, M represents the size of the measurement vector, and N is the number of pixels of the image to be sampled;
the generating an countermeasure network model is as follows:
J(D,G)=min G max D [l rec (G(y))+l reg (G(y))+l GAN (G(y),D)]
wherein G is a generator, D is a discriminator, and the reconstruction loss isl reg For total variation regularization, the countermeasures of the discriminator network are +.>Substituting the reconstruction loss, the total variation regularization and the antagonism loss of the discriminator network into the network model to obtain objective functions respectively optimizing parameters of the generator and the discriminator, wherein the objective functions respectively comprise:
4. A compressed sensing image reconstruction system based on a generation countermeasure network according to claim 3, wherein in the generator training module, an activation function used in training the generator is Selu, and an activation function used in training the discriminator is lrilu.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811404831.XA CN109544656B (en) | 2018-11-23 | 2018-11-23 | Compressed sensing image reconstruction method and system based on generation countermeasure network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811404831.XA CN109544656B (en) | 2018-11-23 | 2018-11-23 | Compressed sensing image reconstruction method and system based on generation countermeasure network |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109544656A CN109544656A (en) | 2019-03-29 |
CN109544656B true CN109544656B (en) | 2023-05-23 |
Family
ID=65850265
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811404831.XA Active CN109544656B (en) | 2018-11-23 | 2018-11-23 | Compressed sensing image reconstruction method and system based on generation countermeasure network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109544656B (en) |
Families Citing this family (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113808062A (en) * | 2019-04-28 | 2021-12-17 | 深圳市商汤科技有限公司 | Image processing method and device |
CN110211045B (en) * | 2019-05-29 | 2022-09-06 | 电子科技大学 | Super-resolution face image reconstruction method based on SRGAN network |
CN110361778B (en) * | 2019-07-04 | 2020-10-13 | 华北电力大学 | Seismic data reconstruction method based on generation countermeasure network |
CN110472667B (en) * | 2019-07-19 | 2024-01-09 | 广东工业大学 | Small target classification method based on deconvolution neural network |
CN110647864A (en) * | 2019-09-30 | 2020-01-03 | 上海依图网络科技有限公司 | Single multi-graph feature recognition method, equipment and medium based on generation countermeasure network |
CN110677671A (en) * | 2019-11-01 | 2020-01-10 | 合肥图鸭信息科技有限公司 | Image compression method and device and terminal equipment |
CN110942097A (en) * | 2019-11-28 | 2020-03-31 | 北京理工大学 | Imaging-free classification method and system based on single-pixel detector |
CN111080727B (en) * | 2019-12-17 | 2023-03-21 | 华中科技大学鄂州工业技术研究院 | Color image reconstruction method and device and image classification method and device |
CN111127454A (en) * | 2019-12-27 | 2020-05-08 | 上海交通大学 | Method and system for generating industrial defect sample based on deep learning |
CN111224992B (en) * | 2020-01-10 | 2022-10-14 | 上海理工大学 | Electric energy quality signal compression reconstruction method based on generation countermeasure network |
CN111382837B (en) * | 2020-02-05 | 2023-07-18 | 鹏城实验室 | Antagonistic sample generation method based on depth product quantization |
CN111461984B (en) * | 2020-03-30 | 2022-06-14 | 华南理工大学 | Load prediction method based on compressed sensing |
CN111711820B (en) * | 2020-05-13 | 2021-03-30 | 哈尔滨工业大学 | Crack image compression sampling method based on generation countermeasure network |
CN111723521B (en) * | 2020-06-12 | 2023-12-08 | 杭州电子科技大学 | Method for recovering limited caliber data into full caliber data based on GAN network |
CN111784581B (en) * | 2020-07-03 | 2024-07-05 | 苏州兴钊防务研究院有限公司 | SAR image super-resolution reconstruction method based on self-normalization generation countermeasure network |
CN112116601B (en) * | 2020-08-18 | 2023-04-28 | 河南大学 | Compressed sensing sampling reconstruction method and system based on generation of countermeasure residual error network |
CN113052925A (en) * | 2021-04-02 | 2021-06-29 | 广东工业大学 | Compressed sensing reconstruction method and system based on deep learning |
CN113361583A (en) * | 2021-06-01 | 2021-09-07 | 珠海大横琴科技发展有限公司 | Countermeasure sample detection method and device |
CN113627538B (en) * | 2021-08-12 | 2024-03-01 | 群联电子股份有限公司 | Method for training asymmetric generation of image generated by countermeasure network and electronic device |
CN113808275B (en) * | 2021-09-24 | 2023-10-13 | 南京信息工程大学 | Single image three-dimensional reconstruction method based on GCN and topology modification |
CN116188259B (en) * | 2022-12-08 | 2024-07-23 | 中国科学院水生生物研究所 | Fish image generation method based on compressed sensing and countermeasure generation |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107784676B (en) * | 2017-09-20 | 2020-06-05 | 中国科学院计算技术研究所 | Compressed sensing measurement matrix optimization method and system based on automatic encoder network |
CN108171762B (en) * | 2017-12-27 | 2021-10-12 | 河海大学常州校区 | Deep learning compressed sensing same-class image rapid reconstruction system and method |
-
2018
- 2018-11-23 CN CN201811404831.XA patent/CN109544656B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN109544656A (en) | 2019-03-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109544656B (en) | Compressed sensing image reconstruction method and system based on generation countermeasure network | |
CN107730451B (en) | Compressed sensing reconstruction method and system based on depth residual error network | |
CN112116601B (en) | Compressed sensing sampling reconstruction method and system based on generation of countermeasure residual error network | |
CN111754403A (en) | Image super-resolution reconstruction method based on residual learning | |
CN108074239B (en) | No-reference image quality objective evaluation method based on prior perception quality characteristic diagram | |
CN109785250B (en) | Image restoration method based on Criminisi algorithm | |
CN102663747B (en) | Stereo image objectivity quality evaluation method based on visual perception | |
US11995801B2 (en) | Image processing method, system, device and storage medium | |
CN110728728B (en) | Compressed sensing network image reconstruction method based on non-local regularization | |
CN110674925B (en) | No-reference VR video quality evaluation method based on 3D convolutional neural network | |
CN107154064B (en) | Natural image compressed sensing method for reconstructing based on depth sparse coding | |
CN112435162B (en) | Terahertz image super-resolution reconstruction method based on complex domain neural network | |
CN107341776A (en) | Single frames super resolution ratio reconstruction method based on sparse coding and combinatorial mapping | |
CN113570658A (en) | Monocular video depth estimation method based on depth convolutional network | |
CN114140442A (en) | Deep learning sparse angle CT reconstruction method based on frequency domain and image domain degradation perception | |
CN108038503A (en) | A kind of woven fabric texture characterizing method based on K-SVD study dictionaries | |
CN115880158A (en) | Blind image super-resolution reconstruction method and system based on variational self-coding | |
CN113723472B (en) | Image classification method based on dynamic filtering constant-variation convolutional network model | |
CN117350373B (en) | Personalized federal aggregation algorithm based on local self-attention mechanism | |
CN110246152A (en) | PIV image processing method and system | |
CN112132924B (en) | CT reconstruction method based on deep neural network | |
CN105608719B (en) | A kind of rapid CT image rebuilding method based on two benches projection adjustment | |
CN105590332A (en) | Rapid algebraic reconstruction technique applied to computed tomography imaging | |
CN110930314B (en) | Image stripe noise suppression method and device, electronic equipment and storage medium | |
CN115984949A (en) | Low-quality face image recognition method and device with attention mechanism |
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 | ||
CB02 | Change of applicant information |
Address after: 210044 No. 219 Ningliu Road, Jiangbei New District, Nanjing City, Jiangsu Province Applicant after: Nanjing University of Information Science and Technology Address before: 210000 Yuting Square, 59 Wangqiao Road, Liuhe District, Nanjing City, Jiangsu Province Applicant before: Nanjing University of Information Science and Technology |
|
CB02 | Change of applicant information | ||
GR01 | Patent grant | ||
GR01 | Patent grant |