CN111444938A - Gas meter character recognition method and system based on width learning algorithm - Google Patents
Gas meter character recognition method and system based on width learning algorithm Download PDFInfo
- Publication number
- CN111444938A CN111444938A CN202010081614.2A CN202010081614A CN111444938A CN 111444938 A CN111444938 A CN 111444938A CN 202010081614 A CN202010081614 A CN 202010081614A CN 111444938 A CN111444938 A CN 111444938A
- Authority
- CN
- China
- Prior art keywords
- matrix
- gas meter
- training
- width learning
- learning algorithm
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/42—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation
- G06V10/422—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation for representing the structure of the pattern or shape of an object therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/513—Sparse representations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/02—Recognising information on displays, dials, clocks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Multimedia (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Medical Informatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Character Discrimination (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a gas meter character recognition method based on a width learning algorithm, which comprises the following steps: completing the training of a width learning algorithm model; acquiring a to-be-identified gas meter character image; carrying out data processing on the gas meter character image to convert the gas meter character image into a digital matrix; and inputting the digital matrix into the trained width learning algorithm model for calculation to obtain the content of the gas meter characters, and completing the gas meter character recognition. The method of the invention adopts a width learning algorithm, can realize the learning of the character counting characteristics of the gas meter, and has high identification accuracy and high identification speed; the method can be realized based on a camera picture, can be based on the existing monitoring equipment or newly-added monitoring equipment, is easy to realize, has low cost and does not involve the reformation of the gas meter; in addition, the method of the present embodiment has an off-line recognition capability without the participation of a cloud.
Description
Technical Field
The invention relates to the field of instrument character recognition, in particular to a gas meter character recognition method and system based on a width learning algorithm.
Background
Along with the development of the society, a large number of water meters, electric meters and gas meters are intelligentized, and manual door-to-door special meter reading is not needed. However, a large number of meters are old meters, and the cost of the intelligent meter is high and the period is long, so that the intelligent meter is still mainly read by manpower, and the unmanned meter reading is realized by a low-cost, quick and simple mode, so that the intelligent meter reading method is a current core appeal.
In the existing gas meter character positioning and recognition method, a traditional image processing method is mainly used for processing a gas meter picture, and the method generally comprises three steps of character positioning, character segmentation and character recognition. The gas meter usually adopts a decimal counting device called a roller counter, and the roller counter has the remarkable characteristics that: the situation that the number display is incomplete and half of the front and back numbers are displayed respectively often occurs, the situation is difficult to process for the traditional character recognition processing system, and the traditional character recognition method is generally complex in process, low in efficiency, poor in system stability and anti-interference performance and low in recognition accuracy.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention aims to provide a gas meter character recognition method and a gas meter character recognition system.
The invention is realized by the following technical scheme:
a gas meter character recognition method based on a width learning algorithm comprises the following steps:
completing the training of a width learning algorithm model;
acquiring a to-be-identified gas meter character image;
carrying out data processing on the gas meter character image to convert the gas meter character image into a digital matrix;
and inputting the digital matrix into the trained width learning algorithm model for calculation to obtain the content of the gas meter characters, and completing the gas meter character recognition.
Preferably, the training of the width learning algorithm model comprises the following steps:
collecting a certain amount of gas meter character images, and constructing training input data for algorithm model training;
processing the training input data to obtain an original input matrix and an original output matrix, and constructing a mapping characteristic point matrix;
constructing an enhanced point matrix by using the mapping characteristic point matrix;
and obtaining a weight matrix from the input layer to the output layer by using a method for solving the pseudo-inverse to complete the training of the algorithm model.
Preferably, the processing the training input data to obtain an original input matrix and an original output matrix, and constructing a mapping feature point matrix specifically includes: and carrying out z-score normalization and sparse representation on the input training data matrix by using a width learning method to generate characteristic nodes and construct a mapping characteristic point matrix.
Preferably, the processing the training data to obtain an original input matrix and an original output matrix, and constructing a mapping feature point matrix, further includes: and performing augmentation processing after z-fraction normalization processing on the input training data matrix by using a width learning method.
Preferably, the constructing an enhanced point matrix by processing the mapping feature point matrix includes: and carrying out normalization and sparse representation on the mapping characteristic point matrix by using a width learning method, generating an enhanced node, and constructing an enhanced point matrix.
Preferably, the collecting a certain amount of gas meter character images and constructing training input data for algorithm model training specifically comprises: dividing a certain amount of gas meter character images into a training data set X Train, a verification data set XINcre and a test data set X test, and performing z-fraction normalization and sparse representation on the training data set X Train, the verification data set X Incre and the test data set Xtest.
Preferably, the processing the training input data to obtain an original input matrix and an original output matrix, and constructing a mapping feature point matrix specifically includes: performing Z-fraction normalization and sparse representation on a training data matrix obtained by an X Train of a training input data set by using a width learning method to generate a characteristic node Zi=φ(XWei+βei) I 1, …, n, and label the feature layer as Zi=[Z1,…,Zi]Wherein W iseiIs a random weight matrix of appropriate dimensions, generated by a random weight matrix ω e in a gaussian distribution, and i represents an iteration amount.
Preferably, the constructing an enhanced point matrix by using the mapping feature point matrix specifically includes: enhanced node H generated by direct calculation based on mapping characteristic point matrixj=φ(ZiWhj+βhj) J-1, …, n, and marks the enhancement layer as Hj=[H1,…,Hj]。
Preferably, the obtaining of the weight matrix from the input layer to the output layer by using the pseudo-inverse computation and the ridge regression algorithm specifically includes: defining the actual content of the gas meter character image as a label vector Y as known data; merging the feature layer and the enhancement layer into A ═ Z | H]The vertical line represents merging the feature layer and the enhancement layer into one line, and calculating the weight W ═ a using the pseudo-inverse and ridge regression algorithm-1Y; after the initial training of the model is completed, the fitting condition and the data generalization capability of the verification data set XINcre and the test data set Xtest verification and debugging model are utilized to complete the training of the model after the expected indexes are reached.
The acquiring of the character image of the gas meter to be identified specifically comprises the following steps: and acquiring an image displayed by characters of the gas meter to be identified through the camera.
The invention also provides a system for realizing the method, which comprises an image acquisition module, a width learning algorithm model training module and an identification module;
the gas meter character image acquisition device is used for acquiring a gas meter character image;
the width learning algorithm model training module is used for realizing width learning algorithm model training and comprises a training input data construction unit, a mapping characteristic point matrix construction unit, an enhancement point matrix construction unit and a weight matrix acquisition unit;
the training input data construction unit is used for constructing training input data for algorithm model training according to a certain amount of gas meter character images collected by the image acquisition module;
the mapping characteristic point matrix construction unit is used for processing the training input data to obtain an original input matrix and an original output matrix and constructing a mapping characteristic point matrix;
the enhanced point matrix constructing unit is used for constructing an enhanced point matrix by using the mapping characteristic point matrix;
the weight matrix acquisition unit is used for acquiring a weight matrix from the input layer to the output layer by using a method for solving the pseudo-inverse to complete the training of the algorithm model;
the identification module comprises a data processing unit and a calculation unit;
the data processing unit is used for processing the gas meter character image to be identified into a digital matrix;
and the calculation unit is used for inputting the digital matrix into the trained width learning algorithm model for calculation to obtain the content of the gas meter characters and finish the gas meter character recognition.
Compared with the prior art, the invention has the following advantages:
the recognition method provided by the invention has the advantages that the model training speed is high, the model can be reconstructed within a few minutes, the recognition accuracy is high, the recognition speed is high, the off-line recognition is supported, the model training can be completed by using a small amount of data and a small amount of time, the performance is excellent, and the implementation difficulty is low.
Drawings
The invention is further described with reference to the following detailed description of embodiments and drawings, in which:
FIG. 1 is a flowchart of a gas meter character recognition method based on a width learning algorithm for completing a width learning algorithm model training according to an embodiment of the present invention;
FIG. 2 is a flow chart of the recognition process of the gas meter character recognition method based on the breadth learning algorithm according to the embodiment of the invention;
FIG. 3 is a schematic block diagram of a gas meter character recognition system based on a breadth learning algorithm according to an embodiment of the present invention.
Detailed Description
The embodiment provides a gas meter character recognition method based on a width learning algorithm, which comprises the following steps:
s1, completing width learning algorithm model training;
s2, acquiring a gas meter character image to be identified;
s3, performing data processing on the gas meter character image to convert the gas meter character image into a digital matrix;
and S4, inputting the number matrix into the trained width learning algorithm model for calculation to obtain the content of the gas meter characters, and completing the gas meter character recognition.
Wherein, S2 to S4 belong to the identification process, as shown in fig. 2.
As shown in fig. 1, the training of the width learning algorithm model in S1 specifically includes the following steps:
s11, acquiring character data of the gas meter, and constructing data input data: collecting a certain amount of gas meter character images through a camera, and constructing training input data for algorithm model training; the specific algorithm is as follows: dividing a certain amount of gas meter character images into a training data set X Train, a verification data set X Incre and a test data set X test, and performing z-fraction normalization and sparse representation on the training data set XTtrain, the verification data set X Incre and the test data set X test;
s12, processing data to obtain an original input matrix and an original output matrix, and constructing a mapping characteristic point matrix: processing the training input data to obtain an original input matrix and an original output matrix, and constructing a mapping characteristic point matrix; more specifically, the training data is sequentially subjected to z-scoresNormalizing and normalizing to obtain an input training data matrix, performing z-fraction normalization and sparse representation on the input training data matrix by using a width learning method, performing augmentation processing after performing z-fraction normalization on the input training data matrix by using the width learning method, generating feature nodes, and constructing a mapping feature point matrix; the specific algorithm is as follows: performing Z-fraction normalization and sparse representation on a training data matrix obtained by an X Train of a training input data set by using a width learning method to generate a characteristic node Zi=φ(XWei+βei) I 1, …, n, and label the feature layer as Zi=[Z1,…,Zi]Wherein W iseiThe random weight matrix is a random weight matrix with proper dimensionality and is generated by a Gaussian-distributed random weight matrix omega e, and i represents an iteration quantity;
s13, constructing an enhanced point matrix by processing the mapping characteristic point matrix: carrying out normalization and sparse representation on the mapping characteristic point matrix by using a width learning method, generating an enhanced node, and constructing an enhanced point matrix; the specific algorithm is as follows: enhanced node H generated by direct calculation based on mapping characteristic point matrixj=φ(ZiWhj+βhj) J-1, …, n, and marks the enhancement layer as Hj=[H1,…,Hj];
S14, obtaining a weight matrix from the input layer to the output layer by a method of solving the pseudo-inverse: defining the actual content of the gas meter character image as a label vector Y as known data; the specific algorithm is as follows: merging the feature layer and the enhancement layer into A ═ Z | H]The vertical line represents merging the feature layer and the enhancement layer into one line, and calculating the weight W ═ a using the pseudo-inverse and ridge regression algorithm- 1Y; after the initial training of the model is completed, the fitting condition and the data generalization capability of the verification model and the debugging model are verified by using a verification data set X Incre and a test data set X test, and the training of the model is completed after the expected indexes are reached;
and S15, finishing the training of the algorithm model.
The gas meter character recognition method is based on a width learning algorithm, and compared with a deep learning algorithm, the width learning algorithm only has 2 layers, namely: the input layer comprises the mapped features, the enhanced nodes and additional added tangent points, the structure is very simple, and the performance can be improved by adding the enhanced nodes; the method of the embodiment adopts a width learning algorithm, can realize the learning of the character counting characteristics of the gas meter, and has high recognition accuracy and high recognition speed; the method can be realized based on a camera picture, can be based on the existing monitoring equipment or newly-added monitoring equipment, is easy to realize, has low cost and does not involve the reformation of the gas meter; in addition, the method of the present embodiment has an off-line recognition capability without the participation of a cloud.
Meanwhile, the present embodiment also provides a system for implementing the above method, and the composition structure of the system is shown in fig. 3, and the system includes five parts, namely, a camera, a data processing module, a training data acquisition and processing module, a width learning algorithm engine, and an output module.
The training data acquisition and processing module is used for importing a gas meter character training data set acquired in advance, carrying out image processing and training a width learning algorithm model to enable the gas meter character training data set to have the recognition capability;
the camera is used for acquiring a character display image of the gas meter;
the data processing module is used for processing the gas meter character display image acquired by the camera, and obtaining an input data matrix X through z-fraction standardization, normalization and augmentation;
the width learning algorithm engine is a core layer of the system, training data are subjected to normalization and sparse representation in the module to generate feature nodes and enhancement nodes, the feature layer and the enhancement layer are constructed, a weight matrix is obtained by solving a pseudo-inverse method, the relation between input and output is found, a gas meter character image acquired by a camera is calculated with the processed input data matrix X and the weight matrix, and then an identification result can be obtained;
and the output module is used for displaying the identification result to the user.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the present invention in any way, and all simple modifications, equivalent variations and modifications made to the above embodiment according to the technical spirit of the present invention still fall within the scope of the technical solution of the present invention.
Claims (11)
1. A gas meter character recognition method based on a width learning algorithm is characterized in that: the method comprises the following steps:
completing the training of a width learning algorithm model;
acquiring a to-be-identified gas meter character image;
carrying out data processing on the gas meter character image to convert the gas meter character image into a digital matrix;
and inputting the digital matrix into the trained width learning algorithm model for calculation to obtain the content of the gas meter characters, and completing the gas meter character recognition.
2. The method of claim 1, wherein: the training of the width learning algorithm model comprises the following steps:
collecting a certain amount of gas meter character images, and constructing training input data for algorithm model training;
processing the training input data to obtain an original input matrix and an original output matrix, and constructing a mapping characteristic point matrix;
constructing an enhanced point matrix by using the mapping characteristic point matrix;
and obtaining a weight matrix from the input layer to the output layer by using a method for solving the pseudo-inverse to complete the training of the algorithm model.
3. The method of claim 2, wherein: the processing of the training input data to obtain an original input matrix and an original output matrix and the construction of a mapping characteristic point matrix specifically comprise: and carrying out z-score normalization and sparse representation on the input training data matrix by using a width learning method to generate characteristic nodes and construct a mapping characteristic point matrix.
4. The method of claim 3, wherein: the processing the training data to obtain an original input matrix and an original output matrix, and constructing a mapping characteristic point matrix, further comprising: and performing augmentation processing after z-fraction normalization processing on the input training data matrix by using a width learning method.
5. The method of claim 2, wherein: the construction of the enhanced point matrix by processing the mapping characteristic point matrix specifically comprises the following steps: and carrying out normalization and sparse representation on the mapping characteristic point matrix by using a width learning method, generating an enhanced node, and constructing an enhanced point matrix.
6. The method according to any one of claims 2 to 5, wherein: the method for collecting a certain amount of gas meter character images and constructing training input data for algorithm model training specifically comprises the following steps: dividing a certain amount of gas meter character images into a training data set X Train, a verification data set X Incre and a test data set X test, and performing z-fraction normalization and sparse representation on the training data set XTtrain, the verification data set X Incre and the test data set X test.
7. The method of claim 6, wherein: the processing of the training input data to obtain an original input matrix and an original output matrix and the construction of a mapping characteristic point matrix specifically comprise: performing Z-fraction normalization and sparse representation on a training data matrix obtained by an X Train of a training input data set by using a width learning method to generate a characteristic node Zi=φ(XWei+βei) I 1, …, and labeled feature layer Zi=[Z1,…,Zi]Wherein W iseiIs a random weight matrix of appropriate dimension, consisting of a Gaussian-distributed random weight matrix omegaeGenerated, i represents the amount of iteration.
8. The method of claim 7, wherein: the constructing of the enhanced point matrix by using the mapping characteristic point matrix specifically comprises: enhanced node H generated by direct calculation based on mapping characteristic point matrixj=φ(ZiWhj+βhj) J-1, …, n, and marks the enhancement layer as Hj=[H1,…,Hj]。
9. The method of claim 8, wherein: the method for obtaining the weight matrix from the input layer to the output layer by using the pseudo-inverse computation and the ridge regression algorithm specifically comprises the following steps: defining the actual content of the gas meter character image as a label vector Y as known data; merging the feature layer and the enhancement layer into A ═ Z | H]The vertical line represents merging the feature layer and the enhancement layer into one line, and calculating the weight W ═ a using the pseudo-inverse and ridge regression algorithm-1Y; after the initial training of the model is completed, the fitting condition and the data generalization capability of the verification model and the debugging model are verified by using the verification data set X Incre and the test data set X test, and the training of the model is completed after the expected indexes are achieved.
10. The method of claim 1, wherein: the acquiring of the character image of the gas meter to be identified specifically comprises the following steps: and acquiring an image displayed by characters of the gas meter to be identified through the camera.
11. A gas meter character recognition system based on a width learning algorithm is characterized in that: the system comprises an image acquisition module, a width learning algorithm model training module and an identification module;
the gas meter character image acquisition device is used for acquiring a gas meter character image;
the width learning algorithm model training module is used for realizing width learning algorithm model training and comprises a training input data construction unit, a mapping characteristic point matrix construction unit, an enhancement point matrix construction unit and a weight matrix acquisition unit;
the training input data construction unit is used for constructing training input data for algorithm model training according to a certain amount of gas meter character images collected by the image acquisition module;
the mapping characteristic point matrix construction unit is used for processing the training input data to obtain an original input matrix and an original output matrix and constructing a mapping characteristic point matrix;
the enhanced point matrix constructing unit is used for constructing an enhanced point matrix by using the mapping characteristic point matrix;
the weight matrix acquisition unit is used for acquiring a weight matrix from the input layer to the output layer by using a method for solving the pseudo-inverse to complete the training of the algorithm model;
the identification module comprises a data processing unit and a calculation unit;
the data processing unit is used for processing the gas meter character image to be identified into a digital matrix;
and the calculation unit is used for inputting the digital matrix into the trained width learning algorithm model for calculation to obtain the content of the gas meter characters and finish the gas meter character recognition.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010081614.2A CN111444938A (en) | 2020-02-06 | 2020-02-06 | Gas meter character recognition method and system based on width learning algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010081614.2A CN111444938A (en) | 2020-02-06 | 2020-02-06 | Gas meter character recognition method and system based on width learning algorithm |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111444938A true CN111444938A (en) | 2020-07-24 |
Family
ID=71650657
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010081614.2A Pending CN111444938A (en) | 2020-02-06 | 2020-02-06 | Gas meter character recognition method and system based on width learning algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111444938A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112949636A (en) * | 2021-03-31 | 2021-06-11 | 上海电机学院 | License plate super-resolution identification method and system and computer readable medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108734301A (en) * | 2017-06-29 | 2018-11-02 | 澳门大学 | A kind of machine learning method and machine learning device |
CN108960422A (en) * | 2018-06-19 | 2018-12-07 | 河南工业大学 | A kind of width learning method based on principal component analysis |
-
2020
- 2020-02-06 CN CN202010081614.2A patent/CN111444938A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108734301A (en) * | 2017-06-29 | 2018-11-02 | 澳门大学 | A kind of machine learning method and machine learning device |
CN108960422A (en) * | 2018-06-19 | 2018-12-07 | 河南工业大学 | A kind of width learning method based on principal component analysis |
Non-Patent Citations (1)
Title |
---|
C.L.PHILIP CHEN 等: "Broad Learning System: An Effective and Efficient Incremental Learning System Without the Need for Deep Architecture", 《IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112949636A (en) * | 2021-03-31 | 2021-06-11 | 上海电机学院 | License plate super-resolution identification method and system and computer readable medium |
CN112949636B (en) * | 2021-03-31 | 2023-05-30 | 上海电机学院 | License plate super-resolution recognition method, system and computer readable medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108960063A (en) | It is a kind of towards event relation coding video in multiple affair natural language description algorithm | |
CN113884290A (en) | Voltage regulator fault diagnosis method based on self-training semi-supervised generation countermeasure network | |
CN106203625A (en) | A kind of deep-neural-network training method based on multiple pre-training | |
CN111369535B (en) | Cell detection method | |
CN114092832A (en) | High-resolution remote sensing image classification method based on parallel hybrid convolutional network | |
CN112070078A (en) | Deep learning-based land utilization classification method and system | |
CN115994325B (en) | Fan icing power generation data enhancement method based on TimeGAN deep learning method | |
CN112884758B (en) | Defect insulator sample generation method and system based on style migration method | |
CN112819853B (en) | Visual odometer method based on semantic priori | |
CN105787895A (en) | Statistical compressed sensing image reconstruction method based on layered Gauss mixing model | |
CN107038730A (en) | The rarefaction representation image rebuilding method being grouped based on Gauss mesostructure block | |
CN115080801A (en) | Cross-modal retrieval method and system based on federal learning and data binary representation | |
CN109461177A (en) | A kind of monocular image depth prediction approach neural network based | |
CN111444938A (en) | Gas meter character recognition method and system based on width learning algorithm | |
CN116385827A (en) | Parameterized face reconstruction model training method and key point tag data generation method | |
CN111444759A (en) | Handwriting recognition method and system based on width learning algorithm | |
CN111310623A (en) | Method for analyzing debris flow sensitivity map based on remote sensing data and machine learning | |
CN113469266A (en) | Electricity stealing behavior detection method based on improved deep convolutional neural network | |
Uittenbogaard et al. | Conditional transfer with dense residual attention: Synthesizing traffic signs from street-view imagery | |
CN112163106A (en) | Second-order similarity perception image Hash code extraction model establishing method and application thereof | |
CN116862080A (en) | Carbon emission prediction method and system based on double-view contrast learning | |
CN115952743A (en) | Multi-source precipitation data collaborative downscaling method and system coupled with random forest and HASM | |
CN116402995A (en) | Lightweight neural network-based ancient architecture point cloud semantic segmentation method and system | |
CN113762481B (en) | Tomographic imaging method and system based on deep learning | |
CN115578511A (en) | Semi-supervised single-view 3D object reconstruction 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 |