CN110739051B - Method for establishing eosinophilic granulocyte proportion model by using nasal polyp pathological picture - Google Patents
Method for establishing eosinophilic granulocyte proportion model by using nasal polyp pathological picture Download PDFInfo
- Publication number
- CN110739051B CN110739051B CN201910951170.0A CN201910951170A CN110739051B CN 110739051 B CN110739051 B CN 110739051B CN 201910951170 A CN201910951170 A CN 201910951170A CN 110739051 B CN110739051 B CN 110739051B
- Authority
- CN
- China
- Prior art keywords
- model
- small
- eosinophil
- proportion
- picture
- 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
Images
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
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/50—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
-
- 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
-
- 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
-
- 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
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Public Health (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Primary Health Care (AREA)
- Epidemiology (AREA)
- Computational Linguistics (AREA)
- Pathology (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Investigating Or Analysing Biological Materials (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a method for establishing an eosinophilic granulocyte proportion model by utilizing a nasal polyp pathological picture, which comprises the following steps: preparing data, namely making nasal polyps into slides, scanning by a digital pathology instrument to obtain a WSI image, and cutting the WSI image to obtain a small pathology image; all the small pathological pictures are divided into training set data and test set data according to a set proportion; establishing an eosinophil proportion model, adopting an inclusion V3 model and training the model on an ImageNet data set to obtain model parameters, removing the last full-connection layer FC of the model, adding a full-connection layer FC with only one neuron in the full-connection layer FC, not adopting any activation function, setting a loss function to adopt Mean Square Error (MSE), and setting a learning rate lr. The method is applied to a pathological auxiliary diagnosis system of chronic nasosinusitis nasal polyp, and the eosinophil granulocyte ratio on a pathological picture is quickly and accurately obtained through learning and training.
Description
Technical Field
The invention relates to the technical field of medical treatment means, in particular to a technology of a method for establishing an eosinophilic granulocyte proportion model by using a nasal polyp pathological picture.
Background
Chronic Rhinosinusitis (CRS) can be clinically classified into non-rhinopolypus (CRSsNP) and rhinopolypus (CRSwNP). Chronic sinusitis with nasal polyps (CRSwNP) is subdivided into two subtypes, eosinophilic nasal polyps (eCRSwNP) and non-eosinophilic nasal polyps (necswnp). Eosinophilic nasal polyps (eCRSwNP) are sensitive to hormonal treatment, whereas non-eosinophilic nasal polyps (necRSWNP) are sensitive to macrolide antibiotics treatment. Clinically, for how to define eosinophilic nasal polyps (eCRSwNP) and non-eosinophilic nasal polyps (neCRSwNP), a pathologist usually randomly takes the average value of the eosinophilic cell ratios under 10 high-power microscopic fields of a patient nasal polyp slide specimen, and takes a diagnosis standard of 10% as a cut-off value to obtain a classification diagnosis (eosinophilic chronic sinusitis is more than or equal to 10% and non-eosinophilic chronic sinusitis is less than 10%).
However, since a slide specimen typically contains hundreds or thousands of fields of view, there can be large sampling differences in the proportion of eosinophils to inflammatory cells at different locations in the specimen. The results of previous researches of the applicant also show that the sampling estimation values of randomly selected 10 visual fields have a plurality of sampling errors with the actual values of the specimen population. The more fields of view a specimen contains, the greater the sampling error will be. In addition, different doctors have different experience or the same doctor has different time, and the sampling estimation values obtained by randomly selecting the visual field are different, that is, the artificial random sampling counting may have measurement deviation.
The proportion of eosinophilic granulocyte to inflammatory cell is counted on the whole slide specimen, so that the diagnosis can be more accurate, sampling errors are avoided, however, a pathologist needs to spend 2-4 hours to completely count one slide specimen, and the time cost is extremely high.
Clinically, the chronic nasosinusitis is diagnosed completely depending on the experience of a pathologist at the present stage, and an objective auxiliary diagnosis system with higher accuracy and timeliness is lacked. The core in this aided diagnosis system is how to obtain the eosinophil fraction of pathological pictures.
Disclosure of Invention
In order to solve the defects of the prior art, the invention provides a method for establishing a eosinophilic granulocyte proportion model by utilizing a nasal polyp pathological picture, which is applied to a chronic nasosinusitis nasal polyp pathological auxiliary diagnosis system and can quickly and accurately obtain the eosinophilic granulocyte proportion value on the pathological picture through learning and training.
The invention is realized by the following technical scheme:
the method for establishing the eosinophilic granulocyte proportion model by utilizing the nasal polyp pathological picture is characterized by comprising the following steps of:
(1) preparing data: the method comprises the steps of manufacturing nasal polyps into slides, scanning the slides through a digital pathology apparatus to obtain a WSI image, and cutting the WSI image to obtain a small pathology image, wherein the WSI is a white slide image for short, namely a digital pathology image;
(2) data marking: taking x as each small pathology picture, counting the number of eosinophils and all inflammatory cells in each small pathology picture, wherein the proportion of eosinophils in inflammatory cells is eosinophil number/all inflammatory cells, counting the eosinophil number N and the non-eosinophil number M, and the proportion S of eosinophils corresponding to each picture is calculated as: N/(N + M), the value range is 0% -100%, wherein, no eosinophil is 0%, all eosinophils are 100%, and average absolute error MAE is adopted to divide all small pathological pictures and corresponding S into training set data and testing set data according to the set proportion;
(3) and (3) modeling algorithm: firstly, an inclusion V3 model under a deep learning keras framework and a model parameter obtained by training the model on an ImageNet data set are adopted, the last full connection layer FC of the model is provided with neurons for classification, a softmax activation function is adopted, the last full connection layer FC of the model is removed, a full connection layer FC is added, the newly added full connection layer FC is provided with only one neuron without any activation function, then the loss function loss of the model of the inclusion V3 is set to be 'MSE', namely the mean square error is adopted, the learning rate lr is set, the open source parameter obtained by training the ImageNet data set is used for carrying out parameter initialization on the inclusion V3 model, finally the parameter of the inclusion V3 model is retrained by the training set data in the step (2), the training times are set to be n rounds, each round is tested by the testing set data, the testing picture is input into the model obtained currently to obtain a prediction value P, calculating the average absolute error MAE (mean absolute error) of the predicted value P and the real label value S of the test data, namely the average value of the absolute values of P and S of each picture; and (3) storing the model parameters corresponding to the minimum MAE in the n rounds so as to obtain the eosinophil proportion model, wherein n is a natural number, when n is 1, the eosinophil proportion model is obtained only by training, and when n is more than 1, training is carried out according to the requirement of the training times.
The set ratio of training set data to test set data is 9:1 or 8: 2.
The obtained eosinophil proportion model is used for giving an eosinophil proportion value in a range of 0-100% for any small pathological picture with the resolution not more than 1024 x 1024 in the nasal polyp digital pathological image.
Integrating eosinophil proportion values of all small pathological pictures on the slide to obtain a final auxiliary diagnosis result of the slide, setting the slide to be composed of N small pathological pictures, wherein N is a natural number, training the eosinophil proportion model respectively to obtain an eosinophil proportion value Si of each small pathological picture, and if i is (1, N), the final diagnosis result of the slide is an average value of N values, and D is ═ Sigma Si/N.
In the modeling algorithm of the step (3), the design is carried out according to the following table:
in the table, conv represents that the convolution kernel is a convolution layer, pool represents a pooling layer, inclusion represents a model module, FC represents a full connection layer, softmax represents a classified activation function, (3 x 3) represents the convolution kernel size, (8 x 8) represents the pooling kernel size, 1000 represents 1000 neurons, and 1 represents 1 neuron.
The invention has the beneficial effects that:
because the eosinophil proportion model is established by utilizing the pathological picture of the nasal polyp, the eosinophil proportion model only needs to be made into a slide, then a WSI picture is obtained by scanning through a digital pathological apparatus, the WSI is cut into pictures to obtain small pathological pictures, the eosinophil proportion value Si of each small pathological picture is obtained by training the eosinophil proportion model, i is (1, N), the final diagnosis result of the slide is the average value of N values, and D is sigma Si/N, so the eosinophil proportion value on the pathological picture can be quickly and accurately obtained through learning and training.
Drawings
FIG. 1 is a flow chart of an embodiment of the present invention;
FIG. 2 is a block diagram of a system according to an embodiment of the present invention;
FIG. 3 is a diagram showing a comparison between the diagnosis accuracy of the present invention;
FIG. 4 is a schematic diagram of the present invention inputting a deep learning prediction model of chronic sinusitis, i.e., an eosinophil proportion model.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the 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.
The invention relates to a method for establishing an eosinophilic granulocyte proportion model by using a nasal polyp pathological picture, which comprises the following steps of:
(1) preparing data: the method comprises the steps of manufacturing nasal polyps into slides, scanning the slides through a digital pathology apparatus to obtain a WSI image, and cutting the WSI image to obtain a small pathology image, wherein the WSI is a white slide image for short, namely a digital pathology image;
in the data preparation of the present invention, firstly, a pathologist delineates a lesion area from a digital pathological image (WSI) by using an open source ASAP (automatic slide analysis platform) software, that is, a delineated area: a region enclosed by the curves; the position of the sketched region is saved by the generated xml-format file.
Then, calling a 'multiresolution image interface' library (https:// githu. com/computationpathlength group/ASAP) of the ASAP itself, and generating a large mask map according to xml, wherein the mask map is a mask schematic diagram, pixel values corresponding to white areas (drawing areas) of the mask map are 1, pixel values corresponding to other black areas of the mask map are 0, and the resolution of the whole mask map is the same as that of a WSI (wireless sensor interface) map, namely a digital pathology slide (whole slide image).
And reading the WSI and the corresponding large mask graph by using an open source "openmask" library (https:// openmask. org /), and cutting the WSI and the corresponding large mask graph according to the set resolution such as 512 x 512 to respectively obtain a small pathology graph and a small mask graph, wherein the positions of the small pathology graph and the small mask graph are in one-to-one correspondence.
And finally, according to the set resolution of the small pictures, cutting the original WSI and the corresponding large mask picture into a plurality of small pathological pictures and small mask pictures by using a sliding window, respectively calculating the pixel average value P (range from 0 to 1) of each small mask picture, setting a threshold value G (range from 0 to 1), only storing the small pathological pictures corresponding to the small pathological pictures with the average value P of the pixels of the small mask pictures being more than or equal to G, and discarding the small pathological pictures corresponding to the small mask pictures with the average value P of the pixels of the small mask pictures being less than G.
In short, it is the cut-map algorithm: the WSI image of the traditional slide after being digitized by the digital pathology scanner usually has 2-4 GB, and the maximum resolution corresponds to about hundred million pixel points. Firstly, a pathologist is requested to draw a region of interest (ROI) through open-source ASAP software to generate an xml file, then a mask (mask) corresponding to the region is generated by using a multiresolution image interface library, then the magnification (level) and picture resolution which need to be cut are determined by using an open-source openslide library, and finally, the picture is automatically cut in the mask region according to the determined level and resolution, for example, thousands to tens of thousands of pictures with 256 × 256 resolutions can be automatically cut when the level of the ROI of a digital pathology WSI is equal to 0.
(2) Data marking: taking x as each small pathology picture, counting the number of eosinophils and all inflammatory cells in each small pathology picture, wherein the proportion of eosinophils in inflammatory cells is eosinophil number/all inflammatory cells, counting the eosinophil number N and the non-eosinophil number M, and the proportion S of eosinophils corresponding to each picture is calculated as: N/(N + M), the value range is 0% to 100%, wherein no eosinophil is 0%, all eosinophils are 100%, the average absolute error MAE is adopted to divide all small pathological images and corresponding S into training set data and test set data according to a set proportion, the most common training set data and test set data are set in a ratio of 9:1 or 8:2, the proportion design is only an example, and the setting is carried out according to the actual situation;
(3) and (3) modeling algorithm: firstly, an inclusion V3 model under a deep learning keras framework is adopted and training is carried out on an ImageNet data set to obtain model parameters, the last layer of the model is a full connection layer FC and has 1000 neurons, a softmax activation function is adopted for classifying 1000 types of nature, then the last layer of the model is removed, a layer of full connection layer FC is added, the newly-added full connection layer FC has only one neuron and does not adopt any activation function, then a loss function loss of the model of the inclusion V3 is set to be MSE, namely mean square error is adopted, a learning rate lr is set to be 0.0008 (empirical value), the inclusion V3 model is subjected to parameter initialization (existing in the keras format of hdf5 in the keras framework) by using open source parameters obtained by training of the ImageNet data set (the data set is an open source non-medical data set), and finally the Inconenet parameters of the inclusion V3 model are retrained by using the data set 3 of the training of the inclusion V model parameters, the training times are set to 50-200 rounds or other design values, each round is tested by using the test set data, a test picture is input into a currently obtained model (parameter) to be predicted to obtain a predicted value P, the predicted value P and a real label value S of the test data are used for calculating an average absolute error MAE (mean absolute error), namely an average value of the absolute values of the predicted value P and the real label value S of each picture. The model parameters corresponding to the minimum MAE in the 100 rounds are stored in hdf5 format, so as to obtain an eosinophil proportion model.
The obtained eosinophil proportion model is used for giving an eosinophil proportion value in a range of 0-100% for any small pathological picture with the resolution not more than 1024 x 1024 in the nasal polyp digital pathological image.
In the modeling algorithm of the step (3), the design is carried out according to the following table:
in the table, conv represents that the convolution kernel is a convolution layer, pool represents a pooling layer, inclusion represents a model module, FC represents a fully connected layer, softmax represents a sorted activation function, (3 x 3) represents the convolution kernel size, (8 x 8) represents the pooling kernel size, 1000 represents 1000 neurons, and 1 represents 1 neuron.
Integrating eosinophil proportion values of all the small pictures on the slide to obtain a final auxiliary diagnosis result of the slide, setting the slide to be composed of N small pathological pictures, wherein N is a natural number, training the eosinophil proportion values respectively by the eosinophil proportion model to obtain an eosinophil proportion value Si of each small pathological picture, and if i is (1, N), the final diagnosis result of the slide is an average value of N values, and D is ═ Sigma Si/N). When N is 1, marking and selecting 1 small pathological picture, actually, a plurality of small pathological pictures are formed, and a more accurate result can be obtained through the average value D.
Therefore, the invention adopts the supervised deep learning regression algorithm to establish an artificial intelligent model, namely an eosinophil proportion model, predicts the proportion of eosinophils in inflammatory cells of each small picture (patch), and finally synthesizes all the small pictures (patch) to obtain the final proportion of eosinophils in inflammatory cells of one digital pathological picture of the patient.
FIG. 1 is a flow chart of an embodiment of the present invention, nasal polyp surgery; manufacturing a pathological slide; a digital pathological scanner is used for digitizing a WSI image, namely a digital pathological slide image; hundreds to thousands of small pictures (patch) with any resolution are obtained after the pictures are cut, namely, the pictures (1), (2) and (3); a supervised deep learning regression model, i.e. a model of the eosinophil proportion; carrying out learning training by using the eosinophil ratio model to obtain predicted values, namely prediction (1), prediction (2) and prediction (3); finally, a digital pathology slide-based diagnosis result is obtained.
FIG. 2 is a block diagram of a system according to an embodiment of the present invention; (1) digital image of nasal polyp slide of chronic nasosinusitis; (2) sketching the digital image of the nasal polyp slide to generate a small picture, namely a small pathological diagram; (3) predicting the eosinophil number ratio of each small picture by using the eosinophil ratio model based on a deep learning quantitative prediction module; (4) and the result output module is used for synthesizing all the small picture prediction results to obtain a final result.
Referring to fig. 3, a comparative graph of diagnostic accuracy in an embodiment of the present invention: the horizontal axis represents the patients (16 persons in total) who participated in the test, the vertical axis represents the ratio of eosinophils (ranging from 0% to 100%), and the dotted line of the red level indicates the clinical diagnostic criteria (eosinophilic sinusitis > 10%, non-eosinophilic sinusitis < non-eosinophilic sinusitis). The true values of the patients' nasal polyp eosinophils are the gold standard (blue line), and the patients can be accurately diagnosed by the present invention (yellow line), while misdiagnosis (red circle) or missed diagnosis (missed diagnosis) is possible when the doctor randomly selects 10 visual field small pictures for diagnosis (black dots).
As described in more detail below: based on the overall statistics of the digital pathological slide, whether the patient is eosinophilic chronic sinusitis or not can be diagnosed more accurately, and the traditional clinical diagnosis method can have missed diagnosis and misdiagnosis (particularly around a 10% diagnosis boundary) because of sampling statistics.
The horizontal axis represents 16 patients, the vertical axis represents eosinophil fraction, the red dotted line represents a 10% diagnostic line, above which the diagnosis is eosinophilic, below which it is non-eosinophilic, the blue line represents the actual eosinophil fraction obtained by the doctor's statistics on the whole slide, and the orange color represents the eosinophil fraction obtained by the deep learning method of the present invention, and each patient corresponds to 50 black dots, each black dot represents 10 randomly selected positions from the doctor for sampling statistics. Examples are: for patient ID 6, the true (blue) value is less than 10% and the value obtained by the method of the present invention is also less than 10%, and therefore all are correctly diagnosed as non-eosinophilic patients, whereas for physician random sampling (black dots), some of the black dots exceed 10%, indicating a misdiagnosis as eosinophilic. Similarly, for patient ID 8, the true (blue) value is greater than 10% and the value obtained by the method of the present invention is also greater than 10%, and therefore correctly diagnosed as eosinophilic, whereas for physician randomized sampling (black dots), there are black dots but less than 10%, indicating a misdiagnosis as non-eosinophilic.
Referring to fig. 4, the eosinophil fraction values of the small picture 1 and the small picture 2 are obtained by inputting the small picture 1 and the small picture 2 into the deep learning prediction model for chronic sinusitis.
In use, the time comparison table for the three methods is as follows, the time required for randomly sampling 10 points by using the existing doctor is about 12 minutes, the time required for manually carrying out complete statistics on the whole slide by using the doctor is about 150 minutes, and the time required for carrying out statistics on the whole slide by using the deep learning method is only about 5 minutes.
The above description is only a preferred embodiment of the present invention, and not intended to limit the scope of the claims, so that equivalent structural changes made by using the description and drawings of the present invention are included in the scope of the present invention.
Claims (4)
1. A method for establishing an eosinophilic granulocyte proportion model by utilizing a nasal polyp pathological picture is characterized by comprising the following steps of:
(1) preparing data: the nasal polyp is made into a slide, then a WSI picture is obtained through scanning of a digital pathology instrument, the WSI picture is cut to obtain a small pathology picture, and the WSI is a white slide image for short, namely a digital pathology image;
in the data preparation, firstly, a diseased region is sketched out through an open source ASAP (automatic site analysis platform) software, namely, a sketched region is drawn out, the position of the sketched region is stored by a generated xml format file, then a 'multiresolution imaging interface' library of the ASAP is called, a large mask map is generated according to xml, the resolution of the whole mask map is the same as that of WSI, then the WSI and the corresponding large mask map are read by the open source 'openside' library, the WSI and the corresponding large mask map are cut according to the set resolution, a small pathology map and a small mask map are respectively obtained, the positions of the small pathology map and the small mask map are in one-to-one correspondence, finally, the original WSI and the corresponding large mask map are cut into a plurality of small pathology pictures and small mask pictures by a sliding window according to the set resolution of the small pictures, the pixel average value P of each small mask map is respectively calculated, setting a threshold value G in a range of 0 to 1, and only storing the small pathology graph corresponding to the small mask graph with the average value P of the pixels of the small mask graph not less than G, and discarding the small pathology graph corresponding to the small mask graph with the average value P of the pixels of the small mask graph less than G;
(2) data marking: taking x as each small pathology picture, counting the number of eosinophils and all inflammatory cells in each small pathology picture, wherein the proportion of eosinophils in inflammatory cells is eosinophil number/all inflammatory cells, counting the eosinophil number N and the non-eosinophil number M, and the proportion S of eosinophils corresponding to each small pathology picture is calculated as: N/(N + M), the value range is 0% -100%, wherein no eosinophil is 0%, all eosinophils are 100%, and all small pathological pictures and corresponding S are divided into training set data and test set data;
(3) and (3) modeling algorithm: firstly, an Inception V3 model under a deep learning keras framework and an ImageNet data set are adopted to be trained to obtain model parameters, the last full-link FC layer of the model is provided with neurons for classification, a softmax activation function is adopted, the last full-link FC layer of the model is removed, a full-link FC layer is added, the added full-link FC layer is provided with only one neuron without any activation function, then the loss function loss of the Inception V3 model is set to be 'MSE', namely the mean square error is adopted, the learning rate lr is set, the open source parameters obtained by training the ImageNet data set are used for carrying out parameter initialization on the Inception V3 model, finally the training set data in the step (2) are used for retraining the parameters of the Inception V3 model, the training times are set to be n rounds, each round is used for testing, the testing set data are input into the model obtained currently to predict values to obtain P, calculating the average absolute error MAE (mean absolute error) of the predicted value P and the real label value S of the test data, namely the average value of the absolute values of P and S of each picture; and storing the corresponding model parameters when the MAE in the n rounds is minimum so as to obtain an eosinophil ratio model, wherein n is a natural number, the eosinophil ratio model is obtained by training once when n is 1, and the eosinophil ratio model is trained according to the requirement of the training times when n is more than 1.
2. The method for establishing an eosinophil proportion model by using a nasal polyp pathological picture according to claim 1, wherein: the set ratio of training set data to test set data is 9:1 or 8: 2.
3. The method for establishing an eosinophil proportion model by using a nasal polyp pathological picture according to claim 1, wherein: the eosinophil proportion value is given by using the obtained eosinophil proportion model to any small pathological picture with the resolution not more than 1024 x 1024 in the nasal polyp digital pathological image, the range is 0-100%,
in the table, conv indicates that the convolution kernel is a convolution layer, pool indicates a pooling layer, inclusion indicates a model module, FC indicates a full-connected layer, softmax indicates a sorted activation function, (3 × 3) indicates the convolution kernel size, (8 × 8) indicates the pooling kernel size, 1000 indicates 1000 neurons, and FC (1)1 indicates 1 neuron.
4. The method for establishing an eosinophil proportion model by using a nasal polyp pathological picture according to claim 1, wherein: integrating eosinophil proportion values of all small pathological pictures on the slide to obtain a final auxiliary diagnosis result of the slide, setting the slide to be composed of N small pathological pictures, wherein N is a natural number, training the eosinophil proportion values respectively by the eosinophil proportion model to obtain an eosinophil proportion value Si of each small pathological picture, and if i is (1, N), the final diagnosis result of the slide is an average value of N values, and D is ═ Sigma Si/N.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910951170.0A CN110739051B (en) | 2019-10-08 | 2019-10-08 | Method for establishing eosinophilic granulocyte proportion model by using nasal polyp pathological picture |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910951170.0A CN110739051B (en) | 2019-10-08 | 2019-10-08 | Method for establishing eosinophilic granulocyte proportion model by using nasal polyp pathological picture |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110739051A CN110739051A (en) | 2020-01-31 |
| CN110739051B true CN110739051B (en) | 2022-06-03 |
Family
ID=69268485
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910951170.0A Active CN110739051B (en) | 2019-10-08 | 2019-10-08 | Method for establishing eosinophilic granulocyte proportion model by using nasal polyp pathological picture |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110739051B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111325220B (en) * | 2020-02-17 | 2023-04-07 | 腾讯科技（深圳）有限公司 | Image generation method, device, equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108564026A (en) * | 2018-04-10 | 2018-09-21 | 复旦大学附属肿瘤医院 | Network establishing method and system for Thyroid Neoplasms smear image classification |
| CN108596046A (en) * | 2018-04-02 | 2018-09-28 | 上海交通大学 | A kind of cell detection method of counting and system based on deep learning |
| CN109190666A (en) * | 2018-07-30 | 2019-01-11 | 北京信息科技大学 | Flowers image classification method based on improved deep neural network |
| CN109740652A (en) * | 2018-12-24 | 2019-05-10 | 深圳大学 | A kind of pathological image classification method and computer equipment |
-
2019
- 2019-10-08 CN CN201910951170.0A patent/CN110739051B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108596046A (en) * | 2018-04-02 | 2018-09-28 | 上海交通大学 | A kind of cell detection method of counting and system based on deep learning |
| CN108564026A (en) * | 2018-04-10 | 2018-09-21 | 复旦大学附属肿瘤医院 | Network establishing method and system for Thyroid Neoplasms smear image classification |
| CN109190666A (en) * | 2018-07-30 | 2019-01-11 | 北京信息科技大学 | Flowers image classification method based on improved deep neural network |
| CN109740652A (en) * | 2018-12-24 | 2019-05-10 | 深圳大学 | A kind of pathological image classification method and computer equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110739051A (en) | 2020-01-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110728666B (en) | Typing method and system for chronic nasosinusitis based on digital pathological slide | |
| US20220343623A1 (en) | Blood smear full-view intelligent analysis method, and blood cell segmentation model and recognition model construction method | |
| CN109670510B (en) | Deep learning-based gastroscope biopsy pathological data screening system | |
| JP7422825B2 (en) | Focus-weighted machine learning classifier error prediction for microscope slide images | |
| CN112101451B (en) | Breast cancer tissue pathological type classification method based on generation of antagonism network screening image block | |
| CN111524137B (en) | Cell identification counting method and device based on image identification and computer equipment | |
| CN108288506A (en) | A kind of cancer pathology aided diagnosis method based on artificial intelligence technology | |
| CN110647875B (en) | Method for segmenting and identifying model structure of blood cells and blood cell identification method | |
| WO2011108551A1 (en) | Diagnostic information distribution device and pathology diagnosis system | |
| Zhang et al. | Automated semantic segmentation of red blood cells for sickle cell disease | |
| CN111291825B (en) | Focus classification model training method, apparatus, computer device and storage medium | |
| CN111488921A (en) | Panoramic digital pathological image intelligent analysis system and method | |
| CN112380900A (en) | Deep learning-based cervical fluid-based cell digital image classification method and system | |
| CN109903282B (en) | Cell counting method, system, device and storage medium | |
| CN112215790A (en) | KI67 index analysis method based on deep learning | |
| CN108305253A (en) | A kind of pathology full slice diagnostic method based on more multiplying power deep learnings | |
| CN115909006B (en) | Mammary tissue image classification method and system based on convolution transducer | |
| CN110838094B (en) | Pathological section dyeing style conversion method and electronic equipment | |
| CN112420170B (en) | Method for improving image classification accuracy of computer aided diagnosis system | |
| CN113052228A (en) | Liver cancer pathological section classification method based on SE-Incepton | |
| CN115661459A (en) | 2D mean teacher model using difference information | |
| CN114445356A (en) | Multi-resolution-based full-field pathological section image tumor rapid positioning method | |
| CN110739051B (en) | Method for establishing eosinophilic granulocyte proportion model by using nasal polyp pathological picture | |
| Huang et al. | HEp-2 cell images classification based on textural and statistic features using self-organizing map | |
| CN114140437A (en) | Fundus hard exudate segmentation method based on deep learning |
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 |