CN111798966A - Scanning method for assisting in identifying digital pathological section based on artificial intelligence - Google Patents
Scanning method for assisting in identifying digital pathological section based on artificial intelligence Download PDFInfo
- Publication number
- CN111798966A CN111798966A CN202010645827.3A CN202010645827A CN111798966A CN 111798966 A CN111798966 A CN 111798966A CN 202010645827 A CN202010645827 A CN 202010645827A CN 111798966 A CN111798966 A CN 111798966A
- Authority
- CN
- China
- Prior art keywords
- digital pathological
- pathological section
- image
- processed
- artificial intelligence
- 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.)
- Withdrawn
Links
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06K—GRAPHICAL DATA READING; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
- G06K17/00—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations
- G06K17/0022—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations arrangements or provisious for transferring data to distant stations, e.g. from a sensing device
- G06K17/0025—Methods or arrangements for effecting co-operative working between equipments covered by two or more of main groups G06K1/00 - G06K15/00, e.g. automatic card files incorporating conveying and reading operations arrangements or provisious for transferring data to distant stations, e.g. from a sensing device the arrangement consisting of a wireless interrogation device in combination with a device for optically marking the record carrier
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H15/00—ICT specially adapted for medical reports, e.g. generation or transmission thereof
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Public Health (AREA)
- Primary Health Care (AREA)
- Medical Informatics (AREA)
- Epidemiology (AREA)
- General Physics & Mathematics (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Computer Networks & Wireless Communication (AREA)
- Investigating Or Analysing Biological Materials (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a scanning method for identifying digital pathological sections based on artificial intelligence assistance, which specifically comprises the following steps: preparing a target area detection model; selecting a digital pathological section to be processed, and collecting information of the digital pathological section to be processed; generating corresponding two-dimensional codes from the information of the digital pathological section to be processed; the scanner scans the digital pathological section to be processed; and detecting the scanned information by using the target area detection model to generate a pathological report and an image. According to the scanning method based on artificial intelligence auxiliary recognition of the digital pathological section, the information of the digital pathological section to be processed and the microscope information of the digital pathological section to be processed are collected, then the information is generated into the corresponding two-dimensional code, the two-dimensional code is attached to the digital pathological section to be processed, when the scanner selects the microscope to scan the digital pathological section to be processed, the appropriate microscope can be automatically selected to be processed according to the information stored in the two-dimensional code, and the scanning mode is simple and quick.
Description
Technical Field
The invention relates to the technical field of scanning methods, in particular to a scanning method for identifying digital pathological sections based on artificial intelligence assistance.
Background
The scanning method for scanning digital pathological sections by the existing scanner has the following defects:
1. the existing scanning method for scanning digital pathological sections by a scanner is to judge how many times of a microscope is needed for the digital pathological sections to be scanned by artificial naked eyes, then to select a corresponding microscope, and then to select a scanning area for scanning by the naked eyes, which consumes manpower and material resources;
2. the existing scanning method for scanning the digital pathological section by the scanner is to comprehensively scan the digital pathological section, and the scanner cannot scan specific areas such as cell areas or stained germ areas on the digital pathological section, so that a plurality of blank areas and irrelevant areas can be scanned, and the working efficiency of the whole scanning process is reduced;
3. in the existing scanning method for scanning the digital pathological section by the scanner, people cannot directly obtain the processing state and corresponding processing information of the section through the digital pathological section.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a scanning method for identifying digital pathological sections based on artificial intelligence assistance, which solves the problems mentioned in the background technology.
In order to achieve the purpose, the invention provides the following technical scheme: a scanning method for identifying digital pathological sections based on artificial intelligence assistance specifically comprises the following steps:
s1: preparing a target area detection model;
s2: selecting a digital pathological section to be processed, and collecting information of the digital pathological section to be processed and microscope information of the digital pathological section to be processed;
s3: generating corresponding two-dimensional codes according to the information of the digital pathological section to be processed and the microscope information of the digital pathological section to be processed, and pasting the two-dimensional codes on the digital pathological section to be processed;
s4: according to the two-dimension code information on the digital pathological section to be processed, the scanner selects a corresponding microscope to scan the digital pathological section to be processed;
s5: the scanning result in step S4 is detected by the target region detection model, and a corresponding result is obtained, and a pathology report and an image are generated.
Preferably, the step of preparing the target region detection model in S1 specifically includes the following steps:
s11: collecting a target digital pathological section, scanning to obtain a pathological image of the target digital pathological section, preprocessing the image, labeling a target area in the image, and then generating a target file;
s12: and selecting a neural network model, and putting the marked image and the target file into the neural network model for deep learning to obtain a target area detection model.
Preferably, in S11, the preprocessing the image includes denoising and cutting the image.
Preferably, the target region in the image in S11 includes a cell region and a stained pathogen region.
Preferably, the target file in S11 includes coordinate information of the target area.
Preferably, the neural network model is composed of a three-channel neural network, each channel is composed of 53 layers of neural networks, each layer of neural network comprises a convolution layer, a pooling layer, a full-link layer and a softmax layer, wherein the final output of the first channel is 13 × 3, the final output of the second channel is 26 × 3, the third channel is 52 × 3, and the information of the three channels is finally integrated, and the final pooling layer, the full-link layer and the softmax layer are added to output the result.
Preferably, a residual error network is attached to each channel of the neural network model, and each residual error module is composed of two convolution layers and a short link.
Preferably, the deep learning includes the steps of:
s121: processing the image, including image denoising and image contrast enhancement;
s122: inputting the data after the image processing in the step S121 into a neural network model, weighting an input value and adding a deviation to each neuron, and then inputting an activation function as the output of the neuron to perform forward propagation to obtain a corresponding data value;
s123: inputting the data value into an error function, comparing the data value with a true value by adding regularization to obtain an error, and judging the recognition degree by judging the error;
s124: determining a gradient vector by back propagation and back derivation, and updating corresponding weights to minimize errors;
s124: and repeating the steps S121-S124 until the average value of the loss errors does not fall any more, completing the deep learning, and storing the weight to obtain the target area detection model.
Preferably, the step S5 of detecting the scanning result in the step S4 by using the target region detection model includes the following steps:
s51: preprocessing an image, including denoising the image and enhancing the contrast of image colors;
s52: inputting the preprocessed data into a target area detection model;
s53: and detecting the area and the position of the target to be detected and outputting a result.
Preferably, the method further comprises step S6: and manufacturing a special two-dimension code identification small program, and checking the processing state of the digital pathological section by scanning the two-dimension code on the digital pathological section.
The invention has the following beneficial effects:
1. according to the scanning method based on artificial intelligence auxiliary recognition of the digital pathological section, the information of the digital pathological section to be processed and the microscope information of the digital pathological section to be processed are collected, then the information is generated into the corresponding two-dimensional code, the two-dimensional code is attached to the digital pathological section to be processed, when the scanner selects the microscope to scan the digital pathological section to be processed, the appropriate microscope can be automatically selected to be processed according to the information stored in the two-dimensional code, and the scanning mode is simple and quick.
2. According to the scanning method based on the artificial intelligence auxiliary recognition digital pathological section, when a digital pathological section to be processed is scanned, a target area detection model used for scanning a cell area of a section or a stained germ area is prepared at first, a scanning result of a scanner scanning the digital pathological section to be processed is detected through the model, and a pathology report and an image of the cell area or the stained germ area on the digital pathological section to be processed can be obtained.
3. According to the scanning method based on artificial intelligence auxiliary identification of the digital pathological section, pathological reports and images of cell areas or stained germ areas on each digital pathological section to be processed are stored in the two-dimensional codes on the section, a special two-dimensional code identification small program is manufactured, the small program can be downloaded to a mobile phone or a computer, and the small program is opened to enable the processing state of the digital pathological section to be processed, such as the pathological reports and the pathological images of the digital pathological section to be processed, to be viewed by scanning the two-dimensional codes on the digital pathological section.
Detailed Description
The technical solutions in the embodiments of the present invention are clearly and completely described below, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
The invention provides a technical scheme that: a scanning method for identifying digital pathological sections based on artificial intelligence assistance specifically comprises the following steps:
s1: preparing a target area detection model;
the preparation of the target area detection model in S1 specifically includes the following steps:
s11: collecting a plurality of target digital pathological sections, scanning to obtain pathological images of the target digital pathological sections, preprocessing the images, wherein the preprocessing of the images comprises denoising and cutting the images, and then marking target areas in the images, wherein the target areas in the images comprise cell areas and stained germ areas, and finally generating a target file which comprises coordinate information of the target areas of the images;
s12: selecting a neural network model, and putting the marked image and the target file into the neural network model for deep learning to obtain a target area detection model;
it should be noted that the neural network model described in S12 is composed of a three-channel neural network, each channel is composed of 53 layers of neural networks, each layer of neural network includes convolutional layers, pooling layers, fully-connected layers and softmax layers, a residual module is attached to each channel of the neural network model, each residual module is composed of two convolutional layers and a short link, the convolution scale of the neural network of each channel is different, the convolution kernel size of the first channel is (5, 5), the convolution kernel size of the second channel is (3, 3), the convolution kernel of the third channel is (1, 1), the final output of the first channel is 13 × 3, the second channel is 26 × 3, the third channel is 52 × 3, the information of the three channels is integrated, and finally, one pooling layer, fully-connected layer and softmax layer are added to output, wherein, the pooling layer is used for down-sampling the image, and a certain area of the image is replaced by the pixel average value of the area, so that the size of the image can be reduced, and the image has translation and rotation invariance; the full connection layer is that each node is connected with all nodes of the previous layer and is used for integrating the extracted characteristics; the Softmax layer obtains the score of each category by weighting and summing the characteristics of the full connection layer, and the Softmax layer is mapped into probability, wherein the Softmax is the ratio of the index of the element to the sum of the indexes of all the elements;
it should be noted that the additional residual error module can enhance the generalization capability of the neural network model, break through the asymmetry of the neural network model, and is more beneficial to the effective update of the parameters of the neural network model, and enhances the working mechanism of gradient propagation: for a multi-layer convolutional network structure, when the input is x, the learned feature is denoted as h (x), we now want to learn the residual f (x) ═ h (x) -x, so that the original learned feature is h (x); this is so because residual learning is easier than original feature direct learning; when the residual error is f (x) equal to 0, the accumulation layer only performs identity mapping at this time, at least the network performance is not reduced, and actually the residual error is not 0, which also enables the accumulation layer to learn a new feature on the basis of the input feature, thereby having better performance;
it should be noted that the neural network model described in S12 has the following features:
a1. concurrency: the input and output can be connected through multiple layers of neurons and the corresponding weight values are stored;
a2. self-learning and organizational ability: the neural network model has strong self-learning ability and can continuously update the weight to adjust the self;
a3. the method has strong robustness and fault tolerance: in the neural network model, the storage of information is distributed on the weight values connected in the whole neural network, so that the neural network model has higher survivability compared with the traditional method, and the damage of a few neurons can not cause the damage of the whole system;
a4. with a high degree of non-linear fit: for more feature vectors, the high-dimensional vectors are difficult to accurately process by the traditional method, and the neural network model can better solve the nonlinear feature vectors according to the adjustment of the weight.
The deep learning described in the step of S12 includes the steps of:
s121: processing the image, including image denoising and image contrast enhancement;
s122: inputting the data after the image processing in the S121 into a neural network model, weighting an input value and adding a deviation to each neuron, and then inputting an activation function as the output of the neuron to forward propagate to obtain a corresponding data value;
s123: inputting the data value into an error function, adding regularization and comparing with a true value to obtain an error, and judging the recognition degree by judging the error, wherein the smaller the error is, the better the error is;
s124: determining a gradient vector by performing back propagation and back derivation on the data value obtained in the step S122, and updating corresponding weight to minimize the error;
s124: repeating the steps S121-S124 until the average value of certain iteration times or loss errors does not fall any more, completing deep learning, and storing the weight to obtain a target area detection model;
s2: selecting a digital pathological section to be processed, and collecting information of the digital pathological section to be processed and microscope information of the digital pathological section to be processed;
s3: generating corresponding two-dimensional codes according to the information of the digital pathological section to be processed and the microscope information of the digital pathological section to be processed, and pasting the two-dimensional codes on the digital pathological section to be processed;
s4: according to the two-dimension code information on the digital pathological section to be processed, the digital pathological section scanner selects a proper microscope and initializes the microscope, and the digital pathological section to be processed is scanned;
s5: detecting the scanning result in the step S4 by using a target area detection model, and generating a pathological report and an image;
it should be noted that: the detecting the scanning result in the step S4 by the target region detection model in the step S5 includes the steps of:
s51: preprocessing an image, including denoising the image and enhancing the contrast of image colors;
s52: inputting the preprocessed data into a target area detection model;
s53: detecting the area and the position of the target to be detected and outputting a result;
s6: and manufacturing a special two-dimension code identification small program, checking the processing state of the digital pathological section by scanning the two-dimension code on the digital pathological section, checking the pathological report if the report is generated, and checking whether the digital pathological section is in the process if the report is not generated.
The working principle of the embodiment is as follows: when a digital pathological section to be processed is scanned, a target area detection model for scanning a section cell area or a dyed germ area is prepared firstly, then information of the digital pathological section to be processed and microscope information thereof are collected, then corresponding two-dimensional codes are generated by the information and are pasted on the digital pathological section to be processed, when a scanner selects the microscope to scan the digital pathological section to be processed, a proper microscope is automatically selected according to the information stored in the two-dimensional codes to process, the scanning result of the scanner scanning the digital pathological section to be processed is detected by the target area detection model, so that pathological reports and images of the cell area or the dyed germ area on the digital pathological section to be processed can be obtained, the pathological reports and the images of the cell area or the dyed germ area on each digital pathological section to be processed are stored in the two-dimensional codes on the sections, a special two-dimensional code identification small program is manufactured, the small program can be downloaded to a mobile phone or a computer, the small program is opened, and the processing state of the digital pathological section to be processed, such as a pathological report and a pathological image of the digital pathological section to be processed, can be checked by scanning the two-dimensional code on the digital pathological section.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (10)
1. A scanning method based on artificial intelligence for assisting in identifying digital pathological sections is characterized by comprising the following steps: the method specifically comprises the following steps:
preparing a target area detection model;
selecting a digital pathological section to be processed, and collecting information of the digital pathological section to be processed;
generating corresponding two-dimensional codes from the information of the digital pathological section to be processed;
the scanner scans the digital pathological section to be processed;
and detecting the scanned information by using the target area detection model to generate a pathological report and an image.
2. The scanning method based on artificial intelligence for assisting in identifying the digital pathological section is characterized in that: the preparation of the target area detection model specifically comprises the following steps:
scanning a target digital pathological section to obtain a pathological image of the digital pathological section, preprocessing the pathological image, labeling a target area in the image, and then generating a target file;
and selecting a neural network model, and putting the marked image and the target file into the neural network model for deep learning to obtain a target area detection model.
3. The scanning method based on artificial intelligence for assisting in identifying the digital pathological section is characterized by comprising the following steps of: the mode of preprocessing the image comprises denoising and cutting the image.
4. The scanning method based on artificial intelligence for assisting in identifying the digital pathological section is characterized by comprising the following steps of: the target area in the image includes a cellular area and a stained pathogen area.
5. The scanning method based on artificial intelligence for assisting in identifying the digital pathological section is characterized by comprising the following steps of: the object file includes coordinate information of the object region.
6. The scanning method based on artificial intelligence for assisting in identifying digital pathological sections according to any one of claims 2 to 5, wherein: the neural network model comprises a three-channel neural network, each channel comprises 53 layers of neural networks, and each layer of neural network comprises a convolutional layer, a pooling layer, a full-link layer and a softmax layer.
7. The scanning method based on artificial intelligence for assisting in identifying the digital pathological section is characterized by comprising the following steps of: a residual error network is added on each channel of the three-channel neural network;
wherein, each residual module consists of two convolution layers and a short link.
8. The scanning method based on artificial intelligence for assisting in identifying the digital pathological section is characterized by comprising the following steps of: the deep learning comprises the following steps:
denoising the image and enhancing the image contrast;
inputting the data after image processing into a neural network model, weighting an input value and adding a deviation to each neuron, and then inputting an activation function as the output of the neuron to forward propagate to obtain a corresponding data value;
inputting the data value into an error function, and comparing the data value with a true value to obtain an error by adding regularization;
and determining a gradient vector by reversely propagating and reversely deriving the data value, updating the weight to minimize the error, and storing the weight to obtain a target area detection model.
9. The scanning method based on artificial intelligence for assisting in identifying the digital pathological section is characterized by comprising the following steps of: the detection of the scanned information by using the target area detection model comprises the following steps:
denoising the image and enhancing the contrast of the image color;
inputting the preprocessed data into a target area detection model;
and detecting the area and the position of the target to be detected and outputting a result.
10. The scanning method based on artificial intelligence for assisting in identifying digital pathological sections according to any one of claims 1-5 or 7-9, wherein: and making a two-dimensional code identification small program for checking the processing state of the digital pathological section.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010645827.3A CN111798966A (en) | 2020-07-07 | 2020-07-07 | Scanning method for assisting in identifying digital pathological section based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010645827.3A CN111798966A (en) | 2020-07-07 | 2020-07-07 | Scanning method for assisting in identifying digital pathological section based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111798966A true CN111798966A (en) | 2020-10-20 |
Family
ID=72810316
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010645827.3A Withdrawn CN111798966A (en) | 2020-07-07 | 2020-07-07 | Scanning method for assisting in identifying digital pathological section based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111798966A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112508010A (en) * | 2020-11-30 | 2021-03-16 | 广州金域医学检验中心有限公司 | Method, system, device and medium for identifying digital pathological section target area |
| CN113378820A (en) * | 2021-07-02 | 2021-09-10 | 深圳市东亿健康服务有限公司 | Method and system for identifying digital pathological section target area |
-
2020
- 2020-07-07 CN CN202010645827.3A patent/CN111798966A/en not_active Withdrawn
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112508010A (en) * | 2020-11-30 | 2021-03-16 | 广州金域医学检验中心有限公司 | Method, system, device and medium for identifying digital pathological section target area |
| WO2022110396A1 (en) * | 2020-11-30 | 2022-06-02 | 广州金域医学检验中心有限公司 | Method, system and device for identifying target area of digital pathology slide, and medium |
| CN113378820A (en) * | 2021-07-02 | 2021-09-10 | 深圳市东亿健康服务有限公司 | Method and system for identifying digital pathological section target area |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111950649B (en) | Attention mechanism and capsule network-based low-illumination image classification method | |
| CN111950453B (en) | Random shape text recognition method based on selective attention mechanism | |
| CN110555399B (en) | Finger vein identification method and device, computer equipment and readable storage medium | |
| Cherny et al. | The recognition of handwritten digits using neural network technology | |
| CN111914727B (en) | Small target human body detection method based on balance sampling and nonlinear feature fusion | |
| CN114627360A (en) | Substation equipment defect identification method based on cascade detection model | |
| CN105447473A (en) | PCANet-CNN-based arbitrary attitude facial expression recognition method | |
| CN113011567B (en) | Training method and device of convolutional neural network model | |
| CN109033107A (en) | Image search method and device, computer equipment and storage medium | |
| CN111798966A (en) | Scanning method for assisting in identifying digital pathological section based on artificial intelligence | |
| CN112395466A (en) | Fraud node identification method based on graph embedded representation and recurrent neural network | |
| CN114898472B (en) | Signature identification method and system based on twin vision transducer network | |
| CN114283120B (en) | Domain-adaptive-based end-to-end multisource heterogeneous remote sensing image change detection method | |
| WO2023116565A1 (en) | Method for intelligently designing network security architecture diagram | |
| CN112465226B (en) | User behavior prediction method based on feature interaction and graph neural network | |
| CN111986193A (en) | Remote sensing image change detection method, electronic equipment and storage medium | |
| CN114419413A (en) | Method for constructing sensing field self-adaptive transformer substation insulator defect detection neural network | |
| CN111768792B (en) | Audio steganalysis method based on convolutional neural network and domain countermeasure learning | |
| CN115984632A (en) | Hyperspectral plastic waste material rapid classification method and device and storage medium | |
| CN114092803A (en) | Cloud detection method and device based on remote sensing image, electronic device and medium | |
| CN110688976A (en) | Store comparison method based on image identification | |
| CN112365451B (en) | Method, device, equipment and computer readable medium for determining image quality grade | |
| CN109271833A (en) | Target identification method, device and electronic equipment based on the sparse self-encoding encoder of stack | |
| CN113486929B (en) | Rock slice image identification method based on residual shrinkage module and attention mechanism | |
| CN114549958B (en) | Night and camouflage target detection method based on context information perception mechanism |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20210526 Address after: No.275, Lane 700, Sichuan North Road, Hongkou District, Shanghai 200080 Applicant after: Shen Yi Address before: Room 120, 1st floor, Lane 16299, Puwei Road, Shanyang Town, Jinshan District, Shanghai Applicant before: Shanghai Changdao Medical Technology Co.,Ltd. |
|
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20201020 |