CN110221346A - A kind of data noise drawing method based on the full convolutional neural networks of residual block - Google Patents
A kind of data noise drawing method based on the full convolutional neural networks of residual block Download PDFInfo
- Publication number
- CN110221346A CN110221346A CN201910608174.9A CN201910608174A CN110221346A CN 110221346 A CN110221346 A CN 110221346A CN 201910608174 A CN201910608174 A CN 201910608174A CN 110221346 A CN110221346 A CN 110221346A
- Authority
- CN
- China
- Prior art keywords
- noise
- residual block
- data
- network
- neural networks
- 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
- 208000037170 Delayed Emergence from Anesthesia Diseases 0.000 title claims abstract description 33
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 15
- 238000012549 training Methods 0.000 claims abstract description 22
- 238000012360 testing method Methods 0.000 claims abstract description 12
- 238000003475 lamination Methods 0.000 claims abstract description 4
- 230000008569 process Effects 0.000 claims description 5
- 238000009826 distribution Methods 0.000 claims description 2
- 230000004927 fusion Effects 0.000 claims 1
- 230000008901 benefit Effects 0.000 abstract description 4
- 238000013135 deep learning Methods 0.000 abstract description 3
- 230000029087 digestion Effects 0.000 abstract description 2
- 238000000605 extraction Methods 0.000 abstract description 2
- 230000000694 effects Effects 0.000 description 6
- 230000006870 function Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 238000013434 data augmentation Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000003825 pressing Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
- G01V1/28—Processing seismic data, e.g. for interpretation or for event detection
- G01V1/36—Effecting static or dynamic corrections on records, e.g. correcting spread; Correlating seismic signals; Eliminating effects of unwanted energy
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V1/00—Seismology; Seismic or acoustic prospecting or detecting
- G01V1/28—Processing seismic data, e.g. for interpretation or for event detection
- G01V1/36—Effecting static or dynamic corrections on records, e.g. correcting spread; Correlating seismic signals; Eliminating effects of unwanted energy
- G01V1/364—Seismic filtering
-
- 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
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V2210/00—Details of seismic processing or analysis
- G01V2210/30—Noise handling
- G01V2210/32—Noise reduction
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Remote Sensing (AREA)
- General Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Environmental & Geological Engineering (AREA)
- Geology (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Geophysics (AREA)
- Theoretical Computer Science (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Error Detection And Correction (AREA)
- Noise Elimination (AREA)
Abstract
The invention discloses a kind of data noise drawing methods based on the full convolutional neural networks of residual block.It is all from same data set using the training set and test set of deep learning method compacting seismic noise, so that the generalization of model is limited.To solve the problems, such as generalization, the design philosophy of network structure is to merge double residual block on the basis of Unet network, to enhance network to the capture ability of random noise.The present invention establishes in coding end to end-decoded network structure, using noisy seismic data as input, the substantive characteristics of random noise is extracted by multiple convolutional layers and residual block, constitutes coding；Decoding is made of multiple warp laminations and residual block again, the output of network is the pressed seismic data of noise.It is compared with current seismic data denoising method; due to having merged double residual block to carry out secondary digestion study to the random noise feature of extraction; the substantive characteristics of noise is learnt more abundant; so having apparent advantage in generalization; not only can effectively Attenuating Random Noise, useful signal can also be protected.
Description
Technical field
The present invention relates to data noise compact technique fields, and in particular to the compacting of earthquake random noise.
Background technique
Traditional seismic data denoising method has the filtering of the domain f-k, the denoising of the domain f-x, wavelet transformation, warp wavelet and discrete remaining
String transformation etc..The above method has been widely used in seismic data denoising field, but still there are noise removal capability deficiencies, destruction
The problems such as useful signal.
In recent years, with the development of depth learning technology, researcher is proposed using depth learning technology to seismic data
The method of denoising.Different from traditional denoising method, deep learning belongs to the scope of statistical learning, and statistical learning can be according to noise
The substantive characteristics of sample learning useful signal and noise, fitting obtains can be by the model of useful signal and noise classification.Due to
This advantage of statistical learning, " Wang Yuqing, Lu Wenkai, Liu Jinlin wait earthquake of the based on data augmentation and CNN to make an uproar at random to document
Acoustic pressure system [J] Chinese Journal of Geophysics, 2019,62 (1): 421-433. " proposes the method based on Unet convolutional neural networks over the ground
The method of the random noise compacting of shake data simultaneously obtains certain effect, but problem is instructed in post-stack seismic data test experiments
Practice collection and test set comes from same data set, the generalization of model is restricted.Document " Ma J.Deep Learning for
Attenuating Random and Coherence Noise Simultaneously[C]//80th EAGE
Conference and Exhibition 2018.2018. " proposition can be pressed based on method of the DnCNN to seismic data noise attenuation
Gaussian noise processed, but the data set for training and test is generated data, is not necessarily obtained in actual seismic data set
Good noise pressing result.Document " Liu J, Lu W, Zhang P.Random Noise Attenuation Using
Convolutional Neural Networks[C]//80th EAGE Conference and Exhibition
2018.2018. the earthquake stochastic noise suppression method based on Unet convolutional neural networks " is proposed, but data set is all from synthesis
Data, generalization is restricted in actual seismic data information.
Summary of the invention
The invention discloses a kind of data noise drawing method based on the full convolutional neural networks of residual block, feature exists
In, comprising the following steps:
Step 1: making the training set and test set of data, it is characterised in that by the noisy data of different noise levels and not
Noisy data choose another block number with training set data different distributions according to as test set as training set；
Step 2: designing a kind of coding and decoded end-to-end network infrastructure, it is characterised in that cataloged procedure is by 5 groups of differences
The double residual block of scale is constituted, and every group of residual block is made of 5 convolutional layers and 1 pond layer；Decoding process and cataloged procedure
Symmetrically, it being made of the double residual block of 4 groups of different scales, every group of residual block is made of 1 warp lamination and 5 convolutional layers, and
Merge the noise characteristic of corresponding coding stage extraction；
Step 3: training network simultaneously saves network model；
Step 4: adjusting parameter selects final ideal model；
Step 5: using the noise for the ideal denoising model compacting data that training obtains, output is the pressed number of noise
According to.
In the step 2, the scale size of 5 groups of double residual blocks of coded portion is respectively 256 × 256,128 × 128,
64 × 64,32 × 32,16 × 16, the scale size of 4 groups of double residual blocks of decoded portion is respectively 32 × 32,64 × 64,128
×128、256×256。
It is compared with current seismic data denoising method, the invention has the following advantages that
(1) RUnet (that is: the full convolutional neural networks of residual block) not only can effectively Attenuating Random Noise, can also protect
Protect useful signal；
(2) the innovation of the invention consists in that compared with Unet convolutional neural networks, due to having merged residual block thus to mentioning
The random noise feature taken has carried out secondary digestion study, the substantive characteristics of noise is learnt it is more abundant, so extensive
Has apparent advantage in property.
Detailed description of the invention
Fig. 1 be the present invention to seismic data random noise compacting flow chart, by containing different noise levels noisyly
Training set of the data as training pattern is shaken, RUnet network, training pattern, according to Y-PSNR and signal-to-noise ratio index are built
Ideal model is selected, test set is inputted into ideal model, output is the pressed data of earthquake random noise；
Fig. 2 is the double residual error block structural diagram merged in inventive network structure, xi-2Indicate the input of residual block；f
(xi-2) indicate the noise characteristic that two layers of convolutional layer extracts；xiIndicate the output characteristic value of residual block, i.e. xi=f (xi-2)+xi-2；
CONV indicates convolutional layer.
Fig. 3 is original Unet convolutional neural networks structure chart；
Fig. 4 is RUnet convolutional neural networks structure chart of the invention, in the basic structure of original Unet convolutional neural networks
On, double residual block is merged, quadratic character study is carried out to noise characteristic, dotted portion is double residual block in figure；
Fig. 5 is that specific implementation of the invention denoises example: a is training set initial data；B is that training set adds data of making an uproar；C is
Denoising result of the RUnet in same data set；D is the noisy data of test set；E is RUnet in different data collection test result；f
For the noise of RUnet removal.
Specific embodiment
In order to effectively remove the random noise in seismic data, set forth herein RUnet convolutional neural networks denoising model, packets
Include following steps.
Step 1: by the seismic data of the Noise of different noise levels and pretreated three-dimensional post-stack seismic data and
As training set, the specific steps are as follows:
(1) 256 are chosen from Parihaka poststack 3-d seismic data set, sampled point is that 256 seismic datas are cut
Piece；
(2) 20%, 25% and 30% gaussian random noise is added to seismic data respectively, and it is corresponding pretreated
Seismic data collectively as training set, wherein plus make an uproar seismic data as input, pretreated seismic data is as label, sample
This amount is 900；
Step 2: network includes a cataloged procedure and a decoding process on the whole.Cataloged procedure is by 5 groups of residual blocks
It constituting, every group of residual block be made of 5 convolutional layers and 1 pond layer, the input data of [256 × 256] dimension is encoded to [16 ×
16] dimensional feature information, convolution kernel are dimensioned to 3 × 3, and step-length is set as 1.It is every to be operated by a residual block, characteristic pattern
The 1/2 of size boil down to last time operation, correspondingly the port number of characteristic pattern is 2 times of last residual block operation, is guaranteed special
Reference breath is not lost.Feature decoding process is made of 4 groups of residual blocks, and each residual block is by 1 warp lamination and 5 convolutional layers
Composition, [16 × 16] the dimensional feature information that will be generated by cataloged procedure are upsampled to the output data of [256 × 256] dimension.With volume
Code process is symmetrical, and every by a residual error operation, the size up-sampling of characteristic pattern is 2 times of last residual error operation, feature
The port number of figure becomes the 1/2 of last residual error operation.The characteristic pattern of coded portion corresponding position is added to decoded portion
In characteristic pattern, in the hope of merging the characteristic information of different scale.Last output is 1 × 1 by a convolution kernel size, step-length 1
Convolutional layer and activation primitive tanh complete, the effect of this layer is similar to full articulamentum；
Step 3: the training set that step 1 is obtained is input in the network model that step 2 is built by queue, using accidentally
Poor backpropagation, and with mean square error loss function come measure network output true value at a distance from label value, with
Machine gradient descent algorithm makes loss function minimum the weight that adjusts between neuron, and by quantitative Y-PSNR,
Signal-to-noise ratio and qualitative visual experience judge that network denoising effect saves each of network model after obtaining optimal denoising effect
Parameter；
Mean square difference formula are as follows:
Y is the true value of network output in formula,For corresponding label value, mean square error is smaller represent true value with
Predicted value is closer, and network is better to the learning effect of training set；
Step 4: choosing several seismic data samples in same 3D data volume as test set, be input to step 3 and obtain
To network model in, and by quantitative Y-PSNR, signal-to-noise ratio and qualitative visual experience judge network denoise effect,
Return step 3 continues trained or adjustment parameter from new training network if it does not meet the requirements, until meeting the requirements, and saves most
Whole ideal network model；
Step 5: using the noise of the network model removal Kerry seismic data cube saved, output is the ground after denoising
Shake data.
Claims (2)
1. the invention discloses a kind of data noise drawing methods based on the full convolutional neural networks of residual block, which is characterized in that
The following steps are included:
Step 1: making the training set and test set of data, it is characterised in that by noisy data of different noise levels and not noisy
Data as training set, choose another block number with training set data different distributions according to as test set；
Step 2: designing a kind of coding and decoded end-to-end network infrastructure, it is characterised in that cataloged procedure is by 5 groups of different scales
Double residual block constitute, every group of residual block be made of 5 convolutional layers and 1 pond layer；Decoding process is symmetrical with cataloged procedure,
It is made of the double residual block of 4 groups of different scales, every group of residual block is made of 1 warp lamination and 5 convolutional layers, and fusion pair
The noise characteristic for answering coding stage to extract；
Step 3: training network simultaneously saves network model；
Step 4: adjusting parameter selects final ideal model；
Step 5: using the noise for the ideal denoising model compacting data that training obtains, output is the pressed data of noise.
2. in claim 1, in the step 2 scale size of 5 groups of double residual blocks of coded portion be respectively 256 × 256,
128 × 128,64 × 64,32 × 32,16 × 16, the scale size of 4 groups of double residual blocks of decoded portion is respectively 32 × 32,
64×64、128×128、256×256。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910608174.9A CN110221346B (en) | 2019-07-08 | 2019-07-08 | Data noise suppression method based on residual block full convolution neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910608174.9A CN110221346B (en) | 2019-07-08 | 2019-07-08 | Data noise suppression method based on residual block full convolution neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110221346A true CN110221346A (en) | 2019-09-10 |
| CN110221346B CN110221346B (en) | 2021-03-09 |
Family
ID=67812768
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910608174.9A Active CN110221346B (en) | 2019-07-08 | 2019-07-08 | Data noise suppression method based on residual block full convolution neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110221346B (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110719240A (en) * | 2019-09-30 | 2020-01-21 | 长沙理工大学 | Channel estimation method and system based on deconvolution and denoising convolution |
| CN110992301A (en) * | 2019-10-14 | 2020-04-10 | 数量级(上海)信息技术有限公司 | Gas contour identification method |
| CN111368710A (en) * | 2020-02-27 | 2020-07-03 | 东北石油大学 | Seismic data random noise suppression method combined with deep learning |
| CN111580161A (en) * | 2020-05-21 | 2020-08-25 | 长江大学 | Earthquake random noise suppression method based on multi-scale convolution self-coding neural network |
| CN111708087A (en) * | 2020-07-09 | 2020-09-25 | 中国科学技术大学 | Method for suppressing seismic data noise based on DnCNN neural network |
| CN111796326A (en) * | 2020-07-07 | 2020-10-20 | 中海石油(中国)有限公司 | Method and system for constructing sequence stratum framework based on Unet network |
| CN111882543A (en) * | 2020-07-29 | 2020-11-03 | 南通大学 | Cigarette filter stick counting method based on AA R2Unet and HMM |
| CN113012050A (en) * | 2019-12-18 | 2021-06-22 | 武汉Tcl集团工业研究院有限公司 | Image processing method and device |
| CN113034373A (en) * | 2019-12-24 | 2021-06-25 | 无锡祥生医疗科技股份有限公司 | Ultrasonic image denoising neural network, denoising method and ultrasonic diagnostic equipment |
| CN113156513A (en) * | 2021-04-14 | 2021-07-23 | 吉林大学 | Convolutional neural network seismic signal denoising method based on attention guidance |
| CN113687414A (en) * | 2021-08-06 | 2021-11-23 | 北京大学 | Data-amplification-based seismic interbed multiple suppression method for convolutional neural network |
| WO2021232653A1 (en) * | 2020-05-21 | 2021-11-25 | 浙江大学 | Pet image reconstruction algorithm combining filtered back-projection algorithm and neural network |
| CN113850269A (en) * | 2021-12-01 | 2021-12-28 | 西南石油大学 | Method for nested connection of residual error network based on multi-branch selective kernel |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108010021A (en) * | 2017-11-30 | 2018-05-08 | 上海联影医疗科技有限公司 | A kind of magic magiscan and method |
| CN108389201A (en) * | 2018-03-16 | 2018-08-10 | 北京推想科技有限公司 | The good pernicious sorting technique of Lung neoplasm based on 3D convolutional neural networks and deep learning |
| CN108550125A (en) * | 2018-04-17 | 2018-09-18 | 南京大学 | A kind of optical distortion modification method based on deep learning |
| US20180314985A1 (en) * | 2017-04-24 | 2018-11-01 | Virginia Tech Intellectual Properties, Inc. | Learning and deploying compression of radio signals |
| CN109840471A (en) * | 2018-12-14 | 2019-06-04 | 天津大学 | A kind of connecting way dividing method based on improvement Unet network model |
| CN109903292A (en) * | 2019-01-24 | 2019-06-18 | 西安交通大学 | A kind of three-dimensional image segmentation method and system based on full convolutional neural networks |
-
2019
- 2019-07-08 CN CN201910608174.9A patent/CN110221346B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180314985A1 (en) * | 2017-04-24 | 2018-11-01 | Virginia Tech Intellectual Properties, Inc. | Learning and deploying compression of radio signals |
| CN108010021A (en) * | 2017-11-30 | 2018-05-08 | 上海联影医疗科技有限公司 | A kind of magic magiscan and method |
| CN108389201A (en) * | 2018-03-16 | 2018-08-10 | 北京推想科技有限公司 | The good pernicious sorting technique of Lung neoplasm based on 3D convolutional neural networks and deep learning |
| CN108550125A (en) * | 2018-04-17 | 2018-09-18 | 南京大学 | A kind of optical distortion modification method based on deep learning |
| CN109840471A (en) * | 2018-12-14 | 2019-06-04 | 天津大学 | A kind of connecting way dividing method based on improvement Unet network model |
| CN109903292A (en) * | 2019-01-24 | 2019-06-18 | 西安交通大学 | A kind of three-dimensional image segmentation method and system based on full convolutional neural networks |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110719240B (en) * | 2019-09-30 | 2022-03-18 | 长沙理工大学 | Channel estimation method and system based on deconvolution and denoising convolution |
| CN110719240A (en) * | 2019-09-30 | 2020-01-21 | 长沙理工大学 | Channel estimation method and system based on deconvolution and denoising convolution |
| CN110992301A (en) * | 2019-10-14 | 2020-04-10 | 数量级(上海)信息技术有限公司 | Gas contour identification method |
| CN113012050A (en) * | 2019-12-18 | 2021-06-22 | 武汉Tcl集团工业研究院有限公司 | Image processing method and device |
| CN113012050B (en) * | 2019-12-18 | 2024-05-24 | 武汉Tcl集团工业研究院有限公司 | Image processing method and device |
| CN113034373A (en) * | 2019-12-24 | 2021-06-25 | 无锡祥生医疗科技股份有限公司 | Ultrasonic image denoising neural network, denoising method and ultrasonic diagnostic equipment |
| CN111368710B (en) * | 2020-02-27 | 2022-03-04 | 东北石油大学 | Seismic data random noise suppression method combined with deep learning |
| CN111368710A (en) * | 2020-02-27 | 2020-07-03 | 东北石油大学 | Seismic data random noise suppression method combined with deep learning |
| WO2021232653A1 (en) * | 2020-05-21 | 2021-11-25 | 浙江大学 | Pet image reconstruction algorithm combining filtered back-projection algorithm and neural network |
| CN111580161A (en) * | 2020-05-21 | 2020-08-25 | 长江大学 | Earthquake random noise suppression method based on multi-scale convolution self-coding neural network |
| CN111796326A (en) * | 2020-07-07 | 2020-10-20 | 中海石油(中国)有限公司 | Method and system for constructing sequence stratum framework based on Unet network |
| CN111796326B (en) * | 2020-07-07 | 2022-11-22 | 中海石油(中国)有限公司 | Method and system for constructing sequence stratum framework based on Unet network |
| CN111708087A (en) * | 2020-07-09 | 2020-09-25 | 中国科学技术大学 | Method for suppressing seismic data noise based on DnCNN neural network |
| CN111882543A (en) * | 2020-07-29 | 2020-11-03 | 南通大学 | Cigarette filter stick counting method based on AA R2Unet and HMM |
| CN111882543B (en) * | 2020-07-29 | 2023-12-26 | 南通大学 | Cigarette filter stick counting method based on AA R2Unet and HMM |
| CN113156513A (en) * | 2021-04-14 | 2021-07-23 | 吉林大学 | Convolutional neural network seismic signal denoising method based on attention guidance |
| CN113156513B (en) * | 2021-04-14 | 2024-01-30 | 吉林大学 | Convolutional neural network seismic signal denoising method based on attention guidance |
| CN113687414A (en) * | 2021-08-06 | 2021-11-23 | 北京大学 | Data-amplification-based seismic interbed multiple suppression method for convolutional neural network |
| CN113687414B (en) * | 2021-08-06 | 2022-07-22 | 北京大学 | Data-augmentation-based seismic interbed multiple suppression method for convolutional neural network |
| CN113850269A (en) * | 2021-12-01 | 2021-12-28 | 西南石油大学 | Method for nested connection of residual error network based on multi-branch selective kernel |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110221346B (en) | 2021-03-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110221346A (en) | A kind of data noise drawing method based on the full convolutional neural networks of residual block | |
| CN109559287A (en) | A kind of semantic image restorative procedure generating confrontation network based on DenseNet | |
| CN111640444A (en) | CNN-based self-adaptive audio steganography method and secret information extraction method | |
| CN108549106B (en) | Aliasing noise drawing method and device | |
| CN112541572A (en) | Residual oil distribution prediction method based on convolutional encoder-decoder network | |
| CN103207409A (en) | Frequency domain full-waveform inversion seismic velocity modeling method | |
| CN101273639A (en) | Method and device for filtering a multidemensional digital signal and associated methods and devices | |
| CN112329780B (en) | Depth image semantic segmentation method based on deep learning | |
| CN113935467A (en) | DAS (data acquisition System) borehole exploration data noise suppression method based on iterative multi-scale attention network | |
| CN104077749A (en) | Seismic data denoising method based on contourlet transformation | |
| CN115436993A (en) | Seismic signal P wave detection method based on PPNet network | |
| CN114757864B (en) | Multi-level fine granularity image generation method based on multi-scale feature decoupling | |
| CN110210498A (en) | Digital image device evidence-obtaining system based on residual error study convolution converged network | |
| CN110968845A (en) | Detection method for LSB steganography based on convolutional neural network generation | |
| CN109508639A (en) | Road scene semantic segmentation method based on multiple dimensioned convolutional neural networks with holes | |
| CN116698890A (en) | Variable-porosity digital core modeling method based on conditional variation self-encoder | |
| CN115877478A (en) | DualGAN-based seismic data high-resolution processing method | |
| CN116318172A (en) | Design simulation software data self-adaptive compression method | |
| CN104780383B (en) | A kind of 3D HEVC multi-resolution video coding methods | |
| CN114460648A (en) | 3D convolutional neural network-based self-supervision 3D seismic data random noise suppression method | |
| CN110533575A (en) | A kind of depth residual error steganalysis method based on isomery core | |
| CN109754024A (en) | Image classification method and device | |
| CN112184552B (en) | Sub-pixel convolution image super-resolution method based on high-frequency feature learning | |
| CN112991169A (en) | Image compression method and system based on image pyramid and generation countermeasure network | |
| CN114200520B (en) | Seismic data denoising method |
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 |