CN116977667B - Tunnel deformation data filling method based on improved GAIN - Google Patents
Tunnel deformation data filling method based on improved GAIN Download PDFInfo
- Publication number
- CN116977667B CN116977667B CN202310958616.9A CN202310958616A CN116977667B CN 116977667 B CN116977667 B CN 116977667B CN 202310958616 A CN202310958616 A CN 202310958616A CN 116977667 B CN116977667 B CN 116977667B
- Authority
- CN
- China
- Prior art keywords
- data
- filling
- tunnel deformation
- autoencoders
- gain
- 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 83
- 238000012544 monitoring process Methods 0.000 claims abstract description 25
- 238000013528 artificial neural network Methods 0.000 claims abstract description 7
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 4
- 230000008569 process Effects 0.000 claims description 18
- 238000012549 training Methods 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 4
- 238000004458 analytical method Methods 0.000 abstract description 4
- 238000012545 processing Methods 0.000 abstract description 3
- 238000004364 calculation method Methods 0.000 abstract description 2
- 238000005429 filling process Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 19
- 230000008859 change Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000012217 deletion Methods 0.000 description 4
- 230000037430 deletion Effects 0.000 description 4
- 230000007547 defect Effects 0.000 description 2
- 238000006073 displacement reaction Methods 0.000 description 2
- 230000002411 adverse Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/72—Data preparation, e.g. statistical preprocessing of image or video features
-
- 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
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Computational Linguistics (AREA)
- Medical Informatics (AREA)
- Databases & Information Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Error Detection And Correction (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a tunnel deformation data filling method based on improved GAIN, which relates to the field of tunnel deformation monitoring data filling, and comprises a method Autoencoders-LSGAIN for carrying out initial interpolation on LSGAIN by a self-encoder, wherein the Autoencoders consists of an encoder and a decoder, the encoder is a function for converting original data into codes, and the encoder comprises a fully connected neural network and a convolutional neural network; the decoder reverts the encoding back to the input, generating an output most similar to the original data. The tunnel deformation data filling method based on the improved GAIN improves the convergence speed of the GIAN method, reduces the calculation time required by filling, enables the filling process to be completed in a shorter time, improves the overall efficiency of data processing, enables the subsequent analysis and modeling work to be spread out more quickly, and has good performance on filling the tunnel deformation monitoring data.
Description
Technical Field
The invention relates to the field of tunnel deformation data filling, in particular to a tunnel deformation data filling method based on improved GAIN.
Background
The processing method of data deletion generally includes deletion method and complement method. The deletion method is to delete the sample or variable with missing data, and this method can maintain the integrity and authenticity of the form, but can lose a great amount of useful information to adversely affect the analysis of the data, and the completion rule is based on the traditional statistical theory such as unconditional mean, median, mode, zero value single value interpolation, etc., to fill in the estimated value at the position of the missing data to maintain the consistency of the data. Yong Huang et al propose to recover health monitoring data using a Bayesian compressed sensing method. Yongchao Yang and Satish Nagarajaiah propose a low-rank matrix alignment method for structural vibration response data of random deletions by sparse recovery based on a minimization theory and by kernel norm minimization. Choi et al used a relational analysis method to estimate the predicted missing strain values using measured strain data during engineering construction. The method is easy to cause sample distortion and data distortion, reduces the robustness of the data, and the interpolated data is easy to be confused with the same value in the original data to destroy the distribution rule of the original data.
Along with the development of the machine learning field, methods such as MF (Miss Forest), K Nearest Neighbor (KNN), de-noising Auto-encoder (DAE) and the like are proposed, and the methods make interpolation values more approximate to true values, but the DAE method has some disadvantages and requires the same type of complete data during training; KNN and MF are not easy to obtain highly correlated information for small sample data, and are easily limited by hardware resources when processing large samples.
The resulting anti-interpolation network (Generative Adversarial Imputation Nets, GAIN) effectively solves the problems associated with the above-described methods. The method for generating the formula does not depend on complete data samples, and has no requirement on the data relationship in the samples participating in training.
Although the GAIN method can fill the missing values of a plurality of features at the same time, there are some defects that the GAIN method defaults to use a value of 0 as an initial value of missing data, but the value of 0 can cause misunderstanding of the model or prejudice of the data when the GAIN method processes the features with very small values, which can cause the GAIN method to influence the authenticity of the tunnel deformation monitoring data when filling the tunnel deformation monitoring data with the value not more than 1. Meanwhile, the GAIN method needs to train the generator and the discriminator at the same time, which results in unstable training of the method, and when the data is completed, the method may bias the generated data set result to a certain single mode in the real data distribution, neglect other modes in the distribution, thereby causing the situation that the generated sample has mode collapse, and cause the quality of the filled data to be reduced when the tunnel deformation monitoring data is filled by the GAIN method, thereby influencing the performance and generalization capability of the final model, and being unfavorable for the subsequent data mining work.
Accordingly, there is a need to provide a tunnel deformation data padding method based on improved GAIN to solve the above-mentioned problems.
Disclosure of Invention
The invention aims to provide a tunnel deformation data filling method based on improved GAIN, which solves the problem that a 0 value is used as an initial interpolation in the GAIN method by adopting a method for carrying out initial interpolation on the GAIN method by a self-encoder, and has good performance on filling tunnel deformation monitoring data when the loss rate of the tunnel deformation monitoring data is not higher than 50 percent.
In order to achieve the above objective, the present invention provides a tunnel deformation data filling method based on improved GAIN, the method of initial interpolation of LSGAIN from encoder Autoencodings-LSGAIN, autoencodings is composed of encoder and decoder, the encoder is a function, the original data is converted into code, the encoder includes fully connected neural network, convolution neural network; the decoder reversely restores the codes back to the input to generate an output most similar to the original data;
the Autoencoders process is as follows, the encoding of the original data X from the input layer to the hidden layer is as shown in formula (1), and the decoding process from the hidden layer to the output layer is as shown in formula (2);
h＝g 1 (X) (1)
wherein X represents the original data,refers to the filled data vector, g 1 Coding process for representing data to hidden layer g 2 Representing the decoding process of the hidden layer to the output layer;
the Autoencoders method performs initial interpolation on GAIN, and comprises the following steps:
S1A: filling the missing tunnel deformation monitoring data by using Autoencoders;
S1B: replacing the missing value of the tunnel deformation monitoring missing data with an Autoencoders filling value;
S1C: filling the data filled by Autoencoders into GAIN;
changing the cross entropy loss function in the GAIN into a least square loss function, and providing LSGAIN based on a least square method; the loss function of the GAIN is shown below,
the loss function of LSGAIN is shown in the following formula;
wherein,refers to a filled data vector, M is a mask vector, represents a random variable of a value {0,1}, H is a random variable depending on a hint matrix, G is a generator for generating data, D is a probability that the data is true or false,representing the expected value.
Preferably, the training process of Autoencoders comprises the following steps:
s1: randomly initializing weights;
s2: calculating a reconstruction error gradient by a back propagation method;
s3: the error is continuously iteratively minimized using an optimization method.
Therefore, the tunnel deformation data filling method based on the improved GAI N has the following beneficial effects:
(1) Compared with the traditional method, the method has certain advantages in filling the defect of the tunnel deformation monitoring data, and the reliability and the authenticity of the tunnel deformation monitoring data are not affected by the tunnel deformation monitoring data filled by the method.
(2) The invention reduces the calculation time required by filling, enables the filling process to be completed in a shorter time, improves the overall efficiency of data processing, and enables the subsequent analysis and modeling work to be developed more quickly.
(3) When the data loss rate is not higher than 50%, the interpolated tunnel deformation monitoring data basically meets the change trend of the tunnel deformation monitoring data, and when the loss rate is higher than 50%, the size of the interpolated tunnel deformation monitoring data is distorted somewhat relative to the original data, but displacement change loss under different working conditions has certain similarity relative to the change of normal data.
The technical scheme of the invention is further described in detail through the drawings and the embodiments.
Drawings
FIG. 1 is a block diagram of a tunnel deformation data padding method based on an improved GAIN of the present invention;
FIG. 2 is a frame diagram of an Autoencoders method of the present invention;
FIG. 3 is a framework diagram of the AE-GAIN method of the present invention;
fig. 4 is a framework diagram of the LSGAIN method of the present invention.
Detailed Description
The technical scheme of the invention is further described below through the attached drawings and the embodiments.
Unless defined otherwise, technical or scientific terms used herein should be given the ordinary meaning as understood by one of ordinary skill in the art to which this invention belongs.
As used herein, the word "comprising" or "comprises" and the like means that elements preceding the word encompass the elements recited after the word, and not exclude the possibility of also encompassing other elements. The terms "inner," "outer," "upper," "lower," and the like are used for convenience in describing and simplifying the description based on the orientation or positional relationship shown in the drawings, and do not denote or imply that the devices or elements referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the invention, but the relative positional relationship may be changed when the absolute position of the object to be described is changed accordingly. In the present invention, unless explicitly specified and limited otherwise, the term "attached" and the like should be construed broadly, and may be, for example, fixedly attached, detachably attached, or integrally formed; can be directly connected or indirectly connected through an intermediate medium, and can be communicated with the inside of two elements or the interaction relationship of the two elements. The specific meaning of the above terms in the present invention can be understood by those of ordinary skill in the art according to the specific circumstances.
As shown in FIG. 1, the invention provides a tunnel deformation data filling method based on improved GAIN, which is a method for carrying out initial interpolation on LSGAIN by a self-encoder, wherein the self-encoder consists of an encoder and a decoder, the encoder is a function for converting original data into codes, and the encoder comprises a fully-connected neural network and a convolutional neural network; the decoder reverts the encoding back to the input, generating an output most similar to the original data.
The training process of Autoencoders comprises the following steps:
s1: randomly initializing weights;
s2: calculating a reconstruction error gradient by a back propagation method;
s3: the error is continuously iteratively minimized using an optimization method.
The Autoencoders process is shown in fig. 2, the encoding of the original data X from the input layer to the hidden layer is shown in formula (1), and the decoding process from the hidden layer to the output layer is shown in formula (2);
h＝g 1 (X) (1)
wherein X represents the original data,refers to the filled data vector, g 1 Coding process for representing data to hidden layer g 2 Representing the decoding process of the hidden layer to the output layer;
as shown in fig. 3, the Autoencoders method performs initial interpolation on the GAIN, including the steps of:
S1A: filling the missing tunnel deformation monitoring data by using Autoencoders;
S1B: replacing the missing value of the tunnel deformation monitoring missing data with an Autoencoders filling value;
S1C: the Autoencoders padded data is filled into GAIN.
Changing the cross entropy loss function in the GAIN into a least square loss function, and providing LSGAIN based on a least square method; the method solves the problems of gradient elimination and mode collapse caused by using a cross entropy loss function in the original method. Meanwhile, more information related to the generator objective function can be found, so that the interpolation method is more accurate, and a frame diagram of the LSGAIN is shown in fig. 4.
The least square loss function is a convex function with good mathematical properties, and the optimal analytic solution of the parameters can be obtained by deriving the loss function, namely, the optimal parameter estimated value can be directly calculated, so that the solving process of the model is simplified; the least squares loss function is an optimization criterion that fits the data by minimizing the difference between the observed and model predictions. Under some assumptions, the least squares solution is unbiased and efficient, which can yield a minimum variance estimate of the parameters, providing a best fit to the dataset; obtaining a global optimal solution: for the least squares loss function, there is a unique globally optimal solution. When the least square method is used for parameter estimation, a globally optimal model fitting result can be ensured to be obtained, and the influence of a locally optimal solution on the result is reduced; the least squares loss function is not only used to fit the data, but also statistical inferences can be made. The significance and reliability of the model are evaluated by carrying out statistical analysis such as hypothesis testing, confidence interval estimation and the like on the model parameters.
The loss function of the GAIN is shown below,
the loss function of LSGAIN is shown in the following formula;
wherein,refers to the filled data vector, M is a mask vector, and represents the random variation of the value {0,1}The quantity, H, is a random variable that depends on the hint matrix, G is a generator used to generate data, D is a probability that the data is true or false,representing the expected value.
Examples
The Autoencoders process is shown in figure 2.
A self-encoder is an unsupervised learning method that automatically learns a feature representation from data and compresses the input data into a low-dimensional encoded form. Autoencoders consist of two parts, encoder and decoder. The encoder is a function that converts raw data into code and is implemented in a variety of ways, including fully connected neural networks, convolutional neural networks. The encoder aims at reducing the data dimension and extracting the main characteristics of the data.
The task of the decoder is to reverse-restore the code back to the input, generating an output most similar to the original data. The decoder is implemented in different ways, such as a back propagation neural network. The object of the decoder is to enable the information obtained from the encoder to recover as much as possible the accuracy of the input data and to make full use of the information in the original data.
Therefore, the method for filling the tunnel deformation data based on the improved GAIN is adopted, when the data loss rate is not higher than 50%, the interpolated tunnel deformation monitoring data basically meets the change trend of the tunnel deformation monitoring data, and when the loss rate is higher than 50%, the size of the interpolated tunnel deformation monitoring data is distorted somewhat relative to the original data, and displacement change loss under different working conditions has certain similarity relative to the change of normal data. In summary, when the loss rate of the tunnel deformation monitoring data is not higher than 50%, the filling of the tunnel deformation monitoring data is well performed.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention and not for limiting it, and although the present invention has been described in detail with reference to the preferred embodiments, it will be understood by those skilled in the art that: the technical scheme of the invention can be modified or replaced by the same, and the modified technical scheme cannot deviate from the spirit and scope of the technical scheme of the invention.
Claims (2)
1. A tunnel deformation data filling method based on improved GAIN is characterized in that: the method for carrying out initial interpolation on LSGAIN by using an automatic encoder comprises an encoder and a decoder, wherein the encoder is a function and converts original data into codes, and the encoder comprises a fully connected neural network and a convolutional neural network; the decoder reversely restores the codes back to the input to generate an output most similar to the original data;
the Autoencoders process is as follows, the encoding of the original data X from the input layer to the hidden layer is as shown in formula (1), and the decoding process from the hidden layer to the output layer is as shown in formula (2);
h＝g 1 (X) (1)
wherein X represents the original data,refers to the filled data vector, g 1 Coding process for representing data to hidden layer g 2 Representing the decoding process of the hidden layer to the output layer;
the Autoencoders method performs initial interpolation on GAIN, and comprises the following steps:
S1A: filling the missing tunnel deformation monitoring data by using Autoencoders;
S1B: replacing the missing value of the tunnel deformation monitoring missing data with an Autoencoders filling value;
S1C: filling the data filled by Autoencoders into GAIN;
changing the cross entropy loss function in the GAIN into a least square loss function, and providing LSGAIN based on a least square method; the loss function of the GAIN is shown below,
the loss function of LSGAIN is shown in the following formula;
wherein,refers to a filled data vector, M is a mask vector, represents a random variable with a value {0,1}, H is a random variable depending on a hint matrix, G is a generator for generating data, D is a probability that the data is true or false, and }>Representing the expected value.
2. The tunnel deformation data padding method based on improved GAIN of claim 1, wherein: the training process of Autoencoders comprises the following steps:
s1: randomly initializing weights;
s2: calculating a reconstruction error gradient by a back propagation method;
s3: the error is continuously iteratively minimized using an optimization method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310958616.9A CN116977667B (en) | 2023-08-01 | 2023-08-01 | Tunnel deformation data filling method based on improved GAIN |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310958616.9A CN116977667B (en) | 2023-08-01 | 2023-08-01 | Tunnel deformation data filling method based on improved GAIN |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116977667A CN116977667A (en) | 2023-10-31 |
| CN116977667B true CN116977667B (en) | 2024-01-26 |
Family
ID=88477868
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310958616.9A Active CN116977667B (en) | 2023-08-01 | 2023-08-01 | Tunnel deformation data filling method based on improved GAIN |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116977667B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109242115A (en) * | 2018-09-17 | 2019-01-18 | 东北大学 | A kind of axial fan hub based on GAN surveys the interpolating method of wind missing data |
| CN114066075A (en) * | 2021-11-22 | 2022-02-18 | 中国农业银行股份有限公司陕西省分行 | Customer loss prediction method based on deep learning |
| CN114494003A (en) * | 2022-04-01 | 2022-05-13 | 华南理工大学 | Ancient character generation method combining shape transformation and texture transformation |
| CN114756727A (en) * | 2022-04-27 | 2022-07-15 | 北京北大软件工程股份有限公司 | Data completion model, data completion obtaining method, data completion obtaining device and medium |
| CN115510042A (en) * | 2022-08-30 | 2022-12-23 | 北京邮电大学 | Power system load data filling method and device based on generation countermeasure network |
| CN116090522A (en) * | 2023-04-07 | 2023-05-09 | 四川大学 | Causal relation discovery method and system for missing data set based on causal feedback |
| CN116166650A (en) * | 2023-02-20 | 2023-05-26 | 重庆大学 | Multisource heterogeneous data cleaning method based on generation countermeasure network |
| WO2023123941A1 (en) * | 2021-12-31 | 2023-07-06 | 深圳前海微众银行股份有限公司 | Data anomaly detection method and apparatus |
| CN116482679A (en) * | 2023-04-26 | 2023-07-25 | 西安电子科技大学 | Method for interpolating radar and photoelectric data missing values by using depth self-encoder |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110320162B (en) * | 2019-05-20 | 2021-04-23 | 广东省智能制造研究所 | Semi-supervised hyperspectral data quantitative analysis method based on generation countermeasure network |
| US20220036199A1 (en) * | 2020-07-31 | 2022-02-03 | VoltSense, Inc. | System and method for predictive foliage impingement and wildfire management using generative adversarial network |
| KR20220038907A (en) * | 2020-09-21 | 2022-03-29 | 삼성에스디에스 주식회사 | Data prediction method based on generative adversarial network and apparatus implementing the same method |
-
2023
- 2023-08-01 CN CN202310958616.9A patent/CN116977667B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109242115A (en) * | 2018-09-17 | 2019-01-18 | 东北大学 | A kind of axial fan hub based on GAN surveys the interpolating method of wind missing data |
| CN114066075A (en) * | 2021-11-22 | 2022-02-18 | 中国农业银行股份有限公司陕西省分行 | Customer loss prediction method based on deep learning |
| WO2023123941A1 (en) * | 2021-12-31 | 2023-07-06 | 深圳前海微众银行股份有限公司 | Data anomaly detection method and apparatus |
| CN114494003A (en) * | 2022-04-01 | 2022-05-13 | 华南理工大学 | Ancient character generation method combining shape transformation and texture transformation |
| CN114756727A (en) * | 2022-04-27 | 2022-07-15 | 北京北大软件工程股份有限公司 | Data completion model, data completion obtaining method, data completion obtaining device and medium |
| CN115510042A (en) * | 2022-08-30 | 2022-12-23 | 北京邮电大学 | Power system load data filling method and device based on generation countermeasure network |
| CN116166650A (en) * | 2023-02-20 | 2023-05-26 | 重庆大学 | Multisource heterogeneous data cleaning method based on generation countermeasure network |
| CN116090522A (en) * | 2023-04-07 | 2023-05-09 | 四川大学 | Causal relation discovery method and system for missing data set based on causal feedback |
| CN116482679A (en) * | 2023-04-26 | 2023-07-25 | 西安电子科技大学 | Method for interpolating radar and photoelectric data missing values by using depth self-encoder |
Non-Patent Citations (2)
| Title |
|---|
| A systematic review of generative adversarial imputation network in missing data imputation;Yuqing Zhang 等;《Neural Computing and Applications》;第35卷;19685–19705 * |
| Least Squares Generative Adversarial Networks;Xudong Mao 等;《2017 IEEE International Conference on Computer Vision》;2813-2821 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116977667A (en) | 2023-10-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113177882B (en) | Single-frame image super-resolution processing method based on diffusion model | |
| CN110248190B (en) | Multilayer residual coefficient image coding method based on compressed sensing | |
| CN111147862B (en) | End-to-end image compression method based on target coding | |
| CN114662788B (en) | Seawater quality three-dimensional time-space sequence multi-parameter accurate prediction method and system | |
| CN114386595B (en) | SAR image compression method based on super prior architecture | |
| CN111402137B (en) | Depth attention coding and decoding single image super-resolution algorithm based on perception loss guidance | |
| CN113328755A (en) | Compressed data transmission method facing edge calculation | |
| CN116600119B (en) | Video encoding method, video decoding method, video encoding device, video decoding device, computer equipment and storage medium | |
| CN116318172A (en) | Design simulation software data self-adaptive compression method | |
| CN110260914B (en) | Engineering safety monitoring system region division method based on measuring point space-time characteristics | |
| CN116977667B (en) | Tunnel deformation data filling method based on improved GAIN | |
| CN114092579B (en) | Point cloud compression method based on implicit neural network | |
| CN114792347A (en) | Image compression method based on multi-scale space and context information fusion | |
| CN110084843A (en) | A kind of method for compressing image based on deep learning applied to furniture 3 D-printing | |
| CN112712855B (en) | Joint training-based clustering method for gene microarray containing deletion value | |
| CN118038032A (en) | Point cloud semantic segmentation model based on super point embedding and clustering and training method thereof | |
| CN117278058A (en) | Data acquisition and processing method for climate financing project | |
| CN111343458B (en) | Sparse gray image coding and decoding method and system based on reconstructed residual | |
| CN116978462A (en) | Method for generating non-natural promoter based on diffusion model | |
| CN115049541B (en) | Reversible gray scale method, system and device based on neural network and image steganography | |
| CN113949880B (en) | Extremely-low-bit-rate man-machine collaborative image coding training method and coding and decoding method | |
| CN116604798A (en) | Method for producing hard PVC microporous foaming product by internal foaming method | |
| CN110349228B (en) | Triangular mesh compression method for data-driven least square prediction | |
| CN114286103A (en) | Density-preserving point cloud compression method based on deep learning | |
| CN116911360B (en) | Method for solving minimum compression rate of semantic information by using neural 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 |