CN107833629A - Aided diagnosis method and system based on deep learning - Google Patents
Aided diagnosis method and system based on deep learning Download PDFInfo
- Publication number
- CN107833629A CN107833629A CN201711010112.5A CN201711010112A CN107833629A CN 107833629 A CN107833629 A CN 107833629A CN 201711010112 A CN201711010112 A CN 201711010112A CN 107833629 A CN107833629 A CN 107833629A
- Authority
- CN
- China
- Prior art keywords
- word
- electronic health
- health record
- key feature
- extraction module
- 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
Abstract
The invention discloses a kind of aided diagnosis method based on deep learning, including:S1, original language material data are carried out with word segmentation processing to establish word insertion inquiry table;S2, by electronic health record data key feature field generate training sample, using word insertion inquiry table digitized, recycle convolutional neural networks generation Accessory Diagnostic Model Based;S3, key feature field is extracted to the electronic health record newly inputted, and inquiry table is embedded in by word and is digitally converted, matched using Accessory Diagnostic Model Based, the diagnostic result of output matching.The present invention also provides a kind of assistant diagnosis system based on deep learning, including corpus data extraction module, word insertion inquiry table structure module, historical electronic medical record data extraction module, new electronic health record data extraction module, word-dividing mode, electronic health record digital module and secondary diagnostic module.Diagnostic result of the present invention in time, accurately, will effectively aid in doctor's quick diagnosis state of an illness.
Description
Technical field
The present invention relates to big data analysis field, and in particular to a kind of aided diagnosis method based on deep learning and is
System.
Background technology
Generally, the doctor of large-scale public hospital needs a large amount of patients with similar symptoms that see and treat patients daily, including warp
The young doctor of deficiency is tested, when the number of seeing and treating patients is excessive, easily occurs that working doctor efficiency is low, even misdiagnosis rate becomes and high asked
Topic.
With the propulsion that China's health medical treatment big data is planned, the medical records system of hospital has been enter into the information age, in hospital
The electronic health record data of magnanimity are saved bit by bit in inside.Detailed note of the patient in hospital diagnosis is contained in these electronic health record data
Record, including Symptoms, the state of an illness and remedy measures etc., making diagnosis for doctor has very high reference value.
Deep learning is quickly grown in recent years, is all achieved in fields such as speech recognition, image recognition, natural language processings
Huge achievement.Thus, it is contemplated that being handled with deep learning electronic health record data, auxiliary diagnosis application is generated,
Unfamiliar doctor is helped to make diagnosis.There is special field (e.g., present illness history, physique in data in view of electronic health record
Inspection result), there is very high reference value for auxiliary diagnosis, on the other hand, traditional aided diagnosis method or system depend on pole
The data of standardization are spent, expense is huge in standardisation process, and because data excessively standardize, in practical application not
Whether " cough one week " and " cough one week " (both at conventional notation) None- identified with doctor's input is same alike result;Together
When, when extracting case history feature, what it was extracted is word one by one for traditional aided diagnosis method or system, rather than entirely
The feature of sentence, when sample is more, feature is too many, extracts completely unrealistic, is not suitable for large-scale data.
The content of the invention
It is an object of the invention to provide a kind of aided diagnosis method and system based on deep learning, it has excellent
Response speed and accuracy rate.
To achieve the above object, the present invention uses following technical scheme:
Aided diagnosis method based on deep learning, comprises the following steps:
S1, by corpus import original language material data, to original language material data carry out word segmentation processing, and establish word insertion look into
Ask table;
Key feature field in S2, extraction electronic health record data, and training sample is generated, use institute's predicate insertion inquiry
Table is digitally converted to training sample, digitlization training sample input convolutional neural networks is trained, generation auxiliary
Diagnostic model;
S3, key feature field is extracted to the electronic health record newly inputted, and generate collection to be predicted, looked into using institute's predicate insertion
Inquiry table is treated forecast set and is digitally converted, and will digitize the collection input to be predicted Accessory Diagnostic Model Based and be matched, defeated
Go out the diagnostic result of matching.
Further, the step S1 is specifically included:
S11, from corpus import original language material data, to original language material data carry out data cleansing;
S12, Chinese word segmentation is carried out to original language material data, obtained word segmentation result is input to term vector model training,
Establish word insertion inquiry table.
Further, the step S2 is specifically included:
S21, by being extracted in electronic health record database the original electron medical record data collection of no mistaken diagnosis is had been acknowledged, from original electricity
Sub- medical record data is concentrated and extracts original medical record data, and data cleansing is carried out to the original medical record data extracted;
S22, feature extraction, key feature field of the extraction doctor in diagnosis are carried out to the medical record data after cleaning;
S23, the key feature field extracted to S22 carry out word segmentation processing, generate training sample;
S24, using institute predicate insertion inquiry table each word in training sample is converted into corresponding term vector;
S25, unified each feature field vector dimension, are finally spliced into the vector representation form of whole piece idagnostic logout, complete
Into the digitlization of training sample;
S26, digitized training sample is input to convolutional neural networks be trained, generate Accessory Diagnostic Model Based.
Further, in the convolutional neural networks in step S26, by convolution kernel with digitizing in training sample one by one
Vector carries out convolution according to a direction, and adds a bias term, and a feature is exported by activation primitive;
Then have, ci=f (w × xi:i+h-1+b);
In formula, ciRepresent new feature;W represents a convolution kernel, represents Spatial Dimension with k, h are included in a window
Idagnostic logout feature unit, then convolution kernel is w ∈ Rhk;xi:i+h-1Represent a key from i-th of key feature field to the i-th+h-1
Vector value between word feature field;B represents a bias term;F is nonlinear function.
Further, step S3 is specifically included:
S31, key feature field is extracted to the electronic health record newly inputted;
S32, word segmentation processing is carried out to the key feature field extracted, obtain collection to be predicted;
S32, using institute predicate insertion inquiry table each word of concentration to be predicted is converted into corresponding term vector;
S33, unified each feature field vector dimension, are finally spliced into the vector representation form of whole piece idagnostic logout, complete
Into the digitlization of collection to be predicted;
S34, by it is digitized it is to be predicted collection be input to the Accessory Diagnostic Model Based, draw the diagnostic result of matching.
Further, the key feature field extracted in step S2 and S3 includes main suit, present illness history and physical examination knot
Fruit.
The present invention also provides a kind of assistant diagnosis system based on deep learning, including:
Corpus data extraction module, the corpus data extraction module imports original language material data from corpus, to original
Beginning corpus data carries out Chinese word segmentation after being cleaned;
Word insertion inquiry table structure module, it is connected with the corpus data extraction module, its be provided with term vector model with
The word segmentation result of the corpus data extraction module is trained, establishes word insertion inquiry table;
Historical electronic medical record data extraction module, it extracts original medical record data from electronic health record database, to extraction
The original medical record data gone out carries out data cleansing, and extracts key feature field of the doctor in diagnosis;
New electronic health record data extraction module, it extracts key feature field from the electronic health record newly inputted;
Word-dividing mode, its respectively with the historical electronic medical record data extraction module and new electronic health record data extraction module
It is connected;Its key feature extracted to the historical electronic medical record data extraction module carries out word segmentation processing, obtains training sample
This;Its key feature field to the new electronic health record data extraction module extraction carries out word segmentation processing, obtains collection to be predicted;
Electronic health record digital module, it is connected with institute's predicate insertion inquiry table structure module and word-dividing mode respectively, its
Institute's predicate insertion inquiry table is called to be digitally converted respectively to the training sample and the collection to be predicted;
Secondary diagnostic module, it is connected with the electronic health record digital module, and it is provided with convolutional neural networks with logarithm
Training sample after word is trained and generates Accessory Diagnostic Model Based;It applies the Accessory Diagnostic Model Based, after digitlization
Collection to be predicted be input, draw diagnostic result and the output of matching.
Further, the historical electronic medical record data extraction module and the new electronic health record data extraction module extraction
Key feature field include main suit, present illness history and physical examination outcomes.
After adopting the above technical scheme, the present invention has the following advantages that compared with background technology:
The present invention is not rely on the standardized data of extreme, but is automatically extracted by deep learning in natural language
Potential feature.It automatically extracts doctor and routinely writes contacting between electronic health record and diagnosis, builds intelligent auxiliary diagnosis model,
Realize the function of auxiliary diagnosis.
What the present invention extracted is the feature of whole sentence, suitable for the extraction of clinical large-scale data, while institute of the present invention
The classification of the convolutional neural networks used, because convolutional neural networks employ the shared network structure of weights, reduce weights
Quantity, reduce the complexity of network, the accuracy rate of classification speed and classification is obtained for very big raising, to reach " auxiliary
Help diagnosis " purpose;
Present invention is particularly suitable for the auxiliary diagnosis of the diseases such as paediatrics.Because the proportion of China pediatrician is low, cause
The task of seeing and treating patients of pediatrician is heavy, is needed simultaneously because the particularity of the constitution of children, during diagnosis double cautious, causes paediatrics to cure
Raw operating pressure is very big.The present invention has efficient diagnosis speed and accurate matching degree, can be provided for doctor effective auxiliary
Diagnosis is helped, high degree avoids mistaken diagnosis, mitigates the operating pressure of doctor.
Brief description of the drawings
Fig. 1 is disclosed aided diagnosis method flow chart;
Fig. 2 is disclosed assistant diagnosis system structured flowchart.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.
Embodiment
As shown in figure 1, the invention discloses a kind of aided diagnosis method based on deep learning, it comprises the following steps:
S1, by corpus import original language material data, to original language material data carry out word segmentation processing, and establish word insertion look into
Ask table;
Key feature field in S2, extraction electronic health record data, and training sample is generated, use institute's predicate insertion inquiry
Table is digitally converted to training sample, digitlization training sample input convolutional neural networks is trained, generation auxiliary
Diagnostic model;
S3, key feature field is extracted to the electronic health record newly inputted, and generate collection to be predicted, looked into using institute's predicate insertion
Inquiry table is treated forecast set and is digitally converted, and will digitize the collection input to be predicted Accessory Diagnostic Model Based and be matched, defeated
Go out the diagnostic result of matching.
Wherein, the step S1 is specifically included:
S11, from corpus import original language material data, to original language material data carry out data cleansing;
S12, Chinese word segmentation is carried out to original language material data, obtained word segmentation result is input to term vector model
(the CBOW models in word2vec) is trained, to train dimension to be embedded in inquiry table for the word of 200 term vector.
Wherein, the step S2 is specifically included:
S21, by being extracted in electronic health record database the original electron medical record data collection of no mistaken diagnosis is had been acknowledged, from original electricity
Sub- medical record data is concentrated and extracts original medical record data, and data cleansing is carried out to the original medical record data extracted;
S22, feature extraction, key feature field (institute of the extraction doctor in diagnosis are carried out to the medical record data after cleaning
Stating key feature field includes main suit, present illness history and physical examination outcomes);
S23, the key feature field extracted to S22 carry out word segmentation processing, generate training sample;
S24, using institute predicate insertion inquiry table term vector conversion is carried out to the word in training sample, by training sample
Each word is converted to corresponding term vector;
S25, unified each feature field vector dimension, the vector representation of generation main suit, present illness history and physical examination outcomes,
The final vector representation form for being spliced into whole piece idagnostic logout, that is, complete the digitlization of training sample;
S26, digitized training sample is input in convolutional neural networks be trained, constantly return adjusting parameter and enter
Row right value update, to reduce error, feature is extracted, generates Accessory Diagnostic Model Based.
In convolutional neural networks in step S26, by convolution kernel with digitlization training sample in one by one vector according to
A direction carries out convolution, and adds a bias term, and a feature is exported by activation primitive.
Then have, ci=f (w × xi:i+h-1+b);In formula, ciRepresent new feature;w∈RhkA convolution kernel is represented, with k generations
Table space dimension, h key feature field is included in a window;The number of key feature field is n, then key feature field
It can be expressed asWhereinAccorded with for attended operation, xi:i+h-1Represent from i-th of pass
Key feature field is to the vector value between the i-th+h-1 key characteristics fields;B represents a bias term;F is one non-linear
Function (such as hyperbolic tangent function (tanh), in the sequence using convolution kernel, such as sequence { x1∶h, x2∶h+1..., xn-h+1:n}
New Feature Mapping can be produced:C=[c1, c2..., cn-h+1], wherein c ∈ Rn-h+1)。
Step S3 is specifically included:
S31, key feature field is extracted to the electronic health record newly inputted;
S32, word segmentation processing is carried out to the key feature field extracted, obtain collection to be predicted;
S32, the word treated using institute predicate insertion inquiry table in forecast set carry out term vector conversion, and each word is converted to
Corresponding term vector;
S33, the dimension by the unified each feature field of sentence vector correlation algorithm, generation main suit, present illness history and physique inspection
Come to an end the vector representation of fruit, is finally spliced into the vector representation form of whole piece idagnostic logout;
S34, by it is digitized it is to be predicted collection be input to the Accessory Diagnostic Model Based, draw the diagnostic result of matching.
The present invention also provides a kind of assistant diagnosis system based on deep learning, for realizing foregoing auxiliary diagnosis side
Method, it includes corpus data extraction module, word insertion inquiry table structure module, historical electronic medical record data extraction module, new electricity
Sub- medical record data extraction module, word-dividing mode, electronic health record digital module and secondary diagnostic module.
The corpus data extraction module imports original language material data from corpus, and original language material data are cleaned
After carry out Chinese word segmentation.
Institute predicate insertion inquiry table structure module is connected with the corpus data extraction module, its provided with term vector model with
The word segmentation result of the corpus data extraction module is trained, establishes word insertion inquiry table.
The historical electronic medical record data extraction module extracts from electronic health record data set and has confirmed that the original of no mistaken diagnosis
Medical record data, data cleansing is carried out to the original medical record data extracted, and extract key feature field of the doctor in diagnosis
(including main suit, present illness history and physical examination outcomes).
The new electronic health record data extraction module extracts key feature field (including master from the electronic health record newly inputted
Tell, present illness history and physical examination outcomes).
The word-dividing mode extracts mould with the historical electronic medical record data extraction module and new electronic health record data respectively
Block is connected.Its key feature extracted to the historical electronic medical record data extraction module carries out word segmentation processing, is trained
Sample;Its key feature field to the new electronic health record data extraction module extraction carries out word segmentation processing, obtains to be predicted
Collection.
The electronic health record digital module is connected with institute's predicate insertion inquiry table structure module and word-dividing mode respectively, its
Institute's predicate insertion inquiry table is called to be digitally converted respectively to the training sample and the collection to be predicted.
The secondary diagnostic module is connected with the electronic health record digital module, and it is provided with convolutional neural networks with logarithm
Training sample after word is trained and generates Accessory Diagnostic Model Based;It applies the Accessory Diagnostic Model Based, after digitlization
Collection to be predicted be input, draw diagnostic result and the output of matching.
The foregoing is only a preferred embodiment of the present invention, but protection scope of the present invention be not limited thereto,
Any one skilled in the art the invention discloses technical scope in, the change or replacement that can readily occur in,
It should all be included within the scope of the present invention.Therefore, protection scope of the present invention should be with scope of the claims
It is defined.
Claims (8)
1. the aided diagnosis method based on deep learning, it is characterised in that comprise the following steps:
S1, by corpus import original language material data, to original language material data carry out word segmentation processing, and establish word insertion inquiry
Table;
Key feature field in S2, extraction electronic health record data, and training sample is generated, use institute's predicate insertion inquiry table pair
Training sample is digitally converted, and digitlization training sample input convolutional neural networks are trained, generate auxiliary diagnosis
Model;
S3, key feature field is extracted to the electronic health record newly inputted, and generate collection to be predicted, use institute's predicate insertion inquiry table
Treat forecast set to be digitally converted, the collection input to be predicted Accessory Diagnostic Model Based will be digitized and matched, output
The diagnostic result matched somebody with somebody.
2. the aided diagnosis method based on deep learning as claimed in claim 1, it is characterised in that the step S1 is specifically wrapped
Include:
S11, from corpus import original language material data, to original language material data carry out data cleansing;
S12, Chinese word segmentation is carried out to corpus data, obtained word segmentation result is input to term vector model training, it is embedding to establish word
Enter inquiry table.
3. the aided diagnosis method based on deep learning as claimed in claim 1, it is characterised in that the step S2 is specifically wrapped
Include:
S21, by being extracted in electronic health record database the original electron medical record data collection of no mistaken diagnosis is had been acknowledged, from original electron disease
Count one by one and extract original medical record data according to concentration, data cleansing is carried out to the original medical record data extracted;
S22, feature extraction, key feature field of the extraction doctor in diagnosis are carried out to the medical record data after cleaning;
S23, the key feature field extracted to S22 carry out word segmentation processing, generate training sample;
S24, using institute predicate insertion inquiry table each word in training sample is converted into corresponding term vector;
S25, unified each feature field vector dimension, the vector representation form of whole piece idagnostic logout is finally spliced into, completes instruction
Practice the digitlization of sample;
S26, digitized training sample is input to convolutional neural networks be trained, generate Accessory Diagnostic Model Based.
4. the aided diagnosis method based on deep learning as claimed in claim 3, it is characterised in that:Convolution in step S26
In neutral net, convolution is carried out according to a direction by convolution kernel and the vector one by one in digitlization training sample, and add
One bias term, a feature is exported by activation primitive;
Then have, ci=f (w × xi∶i+h-1+b);
In formula, ciRepresent new feature;W represents a convolution kernel, represents Spatial Dimension with k, h diagnosis is included in a window
Recording feature unit, then convolution kernel is w ∈ Rhk;xi:i+h-1Represent that a keyword is special from i-th of key feature field to the i-th+h-1
Levy the vector value between field;B represents a bias term;F is nonlinear function.
5. the aided diagnosis method based on deep learning as claimed in claim 1, it is characterised in that step S3 is specifically included:
S31, key feature field is extracted to the electronic health record newly inputted;
S32, word segmentation processing is carried out to the key feature field extracted, obtain collection to be predicted;
S32, using institute predicate insertion inquiry table each word of concentration to be predicted is converted into corresponding term vector;
S33, unified each feature field vector dimension, are finally spliced into the vector representation form of whole piece idagnostic logout, complete to treat
The digitlization of forecast set;
S34, by it is digitized it is to be predicted collection be input to the Accessory Diagnostic Model Based, draw the diagnostic result of matching.
6. the aided diagnosis method based on deep learning as claimed in claim 1, it is characterised in that:Extracted in step S2 and S3
Key feature field include main suit, present illness history and physical examination outcomes.
7. the assistant diagnosis system based on deep learning, it is characterised in that including:
Corpus data extraction module, the corpus data extraction module imports original language material data from corpus, to original language
Material data carry out Chinese word segmentation after carrying out data cleansing;
Word insertion inquiry table structure module, it is connected with the corpus data extraction module, and it is provided with term vector model with to institute
The word segmentation result of predicate material data extraction module is trained, and establishes word insertion inquiry table;
Historical electronic medical record data extraction module, it extracts original medical record data from electronic health record database, to what is extracted
Original medical record data carries out data cleansing, and extracts key feature field of the doctor in diagnosis;
New electronic health record data extraction module, it extracts key feature field from the electronic health record newly inputted;
Word-dividing mode, its respectively with the historical electronic medical record data extraction module and new electronic health record data extraction module phase
Even;Its key feature extracted to the historical electronic medical record data extraction module carries out word segmentation processing, obtains training sample;
Its key feature field to the new electronic health record data extraction module extraction carries out word segmentation processing, obtains collection to be predicted;
Electronic health record digital module, it is connected with institute's predicate insertion inquiry table structure module and word-dividing mode respectively, and it is called
Institute's predicate insertion inquiry table is digitally converted to the training sample and the collection to be predicted respectively;
Secondary diagnostic module, it is connected with the electronic health record digital module, and it is provided with convolutional neural networks with to digitlization
Training sample afterwards is trained and generates Accessory Diagnostic Model Based;It applies the Accessory Diagnostic Model Based, with treating after digitlization
Forecast set is input, draws diagnostic result and the output of matching.
8. the assistant diagnosis system based on deep learning as claimed in claim 7, it is characterised in that:The historical electronic case history
Data extraction module and the key feature field of new electronic health record data extraction module extraction include main suit, present illness history and
Physical examination outcomes.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711010112.5A CN107833629A (en) | 2017-10-25 | 2017-10-25 | Aided diagnosis method and system based on deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711010112.5A CN107833629A (en) | 2017-10-25 | 2017-10-25 | Aided diagnosis method and system based on deep learning |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107833629A true CN107833629A (en) | 2018-03-23 |
Family
ID=61649234
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711010112.5A Pending CN107833629A (en) | 2017-10-25 | 2017-10-25 | Aided diagnosis method and system based on deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107833629A (en) |
Cited By (22)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108766581A (en) * | 2018-05-07 | 2018-11-06 | 上海市公共卫生临床中心 | The key message method for digging and assistant diagnosis system of health medical treatment data |
CN108877914A (en) * | 2018-06-01 | 2018-11-23 | 上海京颐科技股份有限公司 | Behavior monitoring method and device, storage medium, the terminal of patient based on Internet of Things |
CN109036506A (en) * | 2018-07-25 | 2018-12-18 | 平安科技(深圳)有限公司 | Monitoring and managing method, electronic device and the readable storage medium storing program for executing of internet medical treatment interrogation |
CN109087691A (en) * | 2018-08-02 | 2018-12-25 | 科大智能机器人技术有限公司 | A kind of OTC drugs recommender system and recommended method based on deep learning |
CN109192299A (en) * | 2018-08-13 | 2019-01-11 | 中国科学院计算技术研究所 | A kind of medical analysis auxiliary system based on convolutional neural networks |
CN109599185A (en) * | 2018-11-14 | 2019-04-09 | 金色熊猫有限公司 | Disease data processing method, device, electronic equipment and computer-readable medium |
CN109785928A (en) * | 2018-12-25 | 2019-05-21 | 平安科技(深圳)有限公司 | Diagnosis and treatment proposal recommending method, device and storage medium |
CN109920501A (en) * | 2019-01-24 | 2019-06-21 | 西安交通大学 | Electronic health record classification method and system based on convolutional neural networks and Active Learning |
CN109935336A (en) * | 2019-01-15 | 2019-06-25 | 北京思普科软件股份有限公司 | A kind of the intelligent auxiliary diagnosis method and diagnostic system of children's division of respiratory disease disease |
CN109949929A (en) * | 2019-03-19 | 2019-06-28 | 挂号网(杭州)科技有限公司 | A kind of assistant diagnosis system based on the extensive case history of deep learning |
CN110164519A (en) * | 2019-05-06 | 2019-08-23 | 北京工业大学 | A kind of classification method for being used to handle electronic health record blended data based on many intelligence networks |
CN110276749A (en) * | 2019-06-14 | 2019-09-24 | 辽宁万象联合医疗科技有限公司 | Children penetrate the quality control artificial intelligence system and its quality control method of piece and diagnosis |
WO2020037454A1 (en) * | 2018-08-20 | 2020-02-27 | 深圳市全息医疗科技有限公司 | Smart auxiliary diagnosis and treatment system and method |
CN111028939A (en) * | 2019-11-15 | 2020-04-17 | 华南理工大学 | Multigroup intelligent diagnosis system based on deep learning |
CN111192680A (en) * | 2019-12-25 | 2020-05-22 | 山东众阳健康科技集团有限公司 | Intelligent auxiliary diagnosis method based on deep learning and collective classification |
CN111199797A (en) * | 2019-12-31 | 2020-05-26 | 中国中医科学院中医药信息研究所 | Auxiliary diagnosis model establishing and auxiliary diagnosis method and device |
WO2020133358A1 (en) * | 2018-12-29 | 2020-07-02 | 深圳市优必选科技有限公司 | Chat corpus cleaning method, apparatus, computer device and storage medium |
CN111524570A (en) * | 2020-05-06 | 2020-08-11 | 万达信息股份有限公司 | Ultrasonic follow-up patient screening method based on machine learning |
CN111599462A (en) * | 2020-05-09 | 2020-08-28 | 吾征智能技术(北京)有限公司 | Intelligent body abnormal smell screening system based on cognitive learning |
CN111739643A (en) * | 2020-08-20 | 2020-10-02 | 耀方信息技术(上海)有限公司 | Health risk assessment system |
CN111883251A (en) * | 2020-07-28 | 2020-11-03 | 平安科技(深圳)有限公司 | Medical misdiagnosis detection method and device, electronic equipment and storage medium |
CN112259229A (en) * | 2020-06-10 | 2021-01-22 | 北京小白世纪网络科技有限公司 | Method, system and equipment for realizing traditional Chinese medicine diagnosis and treatment based on neural network |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106339591A (en) * | 2016-08-25 | 2017-01-18 | 汤平 | Breast cancer prevention self-service health cloud service system based on deep convolutional neural network |
CN106778014A (en) * | 2016-12-29 | 2017-05-31 | 浙江大学 | A kind of risk Forecasting Methodology based on Recognition with Recurrent Neural Network |
CN106874643A (en) * | 2016-12-27 | 2017-06-20 | 中国科学院自动化研究所 | Build the method and system that knowledge base realizes assisting in diagnosis and treatment automatically based on term vector |
US9754220B1 (en) * | 2016-07-01 | 2017-09-05 | Intraspexion Inc. | Using classified text and deep learning algorithms to identify medical risk and provide early warning |
-
2017
- 2017-10-25 CN CN201711010112.5A patent/CN107833629A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9754220B1 (en) * | 2016-07-01 | 2017-09-05 | Intraspexion Inc. | Using classified text and deep learning algorithms to identify medical risk and provide early warning |
CN106339591A (en) * | 2016-08-25 | 2017-01-18 | 汤平 | Breast cancer prevention self-service health cloud service system based on deep convolutional neural network |
CN106874643A (en) * | 2016-12-27 | 2017-06-20 | 中国科学院自动化研究所 | Build the method and system that knowledge base realizes assisting in diagnosis and treatment automatically based on term vector |
CN106778014A (en) * | 2016-12-29 | 2017-05-31 | 浙江大学 | A kind of risk Forecasting Methodology based on Recognition with Recurrent Neural Network |
Non-Patent Citations (1)
Title |
---|
李超: "智能疾病导诊及医疗问答方法研究与应用", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (31)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108766581A (en) * | 2018-05-07 | 2018-11-06 | 上海市公共卫生临床中心 | The key message method for digging and assistant diagnosis system of health medical treatment data |
CN108877914A (en) * | 2018-06-01 | 2018-11-23 | 上海京颐科技股份有限公司 | Behavior monitoring method and device, storage medium, the terminal of patient based on Internet of Things |
CN109036506A (en) * | 2018-07-25 | 2018-12-18 | 平安科技(深圳)有限公司 | Monitoring and managing method, electronic device and the readable storage medium storing program for executing of internet medical treatment interrogation |
CN109087691A (en) * | 2018-08-02 | 2018-12-25 | 科大智能机器人技术有限公司 | A kind of OTC drugs recommender system and recommended method based on deep learning |
CN109192299A (en) * | 2018-08-13 | 2019-01-11 | 中国科学院计算技术研究所 | A kind of medical analysis auxiliary system based on convolutional neural networks |
WO2020037454A1 (en) * | 2018-08-20 | 2020-02-27 | 深圳市全息医疗科技有限公司 | Smart auxiliary diagnosis and treatment system and method |
CN109599185A (en) * | 2018-11-14 | 2019-04-09 | 金色熊猫有限公司 | Disease data processing method, device, electronic equipment and computer-readable medium |
CN109599185B (en) * | 2018-11-14 | 2021-05-25 | 金色熊猫有限公司 | Disease data processing method and device, electronic equipment and computer readable medium |
CN109785928A (en) * | 2018-12-25 | 2019-05-21 | 平安科技(深圳)有限公司 | Diagnosis and treatment proposal recommending method, device and storage medium |
WO2020133358A1 (en) * | 2018-12-29 | 2020-07-02 | 深圳市优必选科技有限公司 | Chat corpus cleaning method, apparatus, computer device and storage medium |
CN109935336B (en) * | 2019-01-15 | 2023-10-31 | 北京思普科软件股份有限公司 | Intelligent auxiliary diagnosis system for respiratory diseases of children |
CN109935336A (en) * | 2019-01-15 | 2019-06-25 | 北京思普科软件股份有限公司 | A kind of the intelligent auxiliary diagnosis method and diagnostic system of children's division of respiratory disease disease |
CN109920501B (en) * | 2019-01-24 | 2021-04-20 | 西安交通大学 | Electronic medical record classification method and system based on convolutional neural network and active learning |
CN109920501A (en) * | 2019-01-24 | 2019-06-21 | 西安交通大学 | Electronic health record classification method and system based on convolutional neural networks and Active Learning |
CN109949929A (en) * | 2019-03-19 | 2019-06-28 | 挂号网(杭州)科技有限公司 | A kind of assistant diagnosis system based on the extensive case history of deep learning |
CN110164519A (en) * | 2019-05-06 | 2019-08-23 | 北京工业大学 | A kind of classification method for being used to handle electronic health record blended data based on many intelligence networks |
CN110276749A (en) * | 2019-06-14 | 2019-09-24 | 辽宁万象联合医疗科技有限公司 | Children penetrate the quality control artificial intelligence system and its quality control method of piece and diagnosis |
CN110276749B (en) * | 2019-06-14 | 2022-04-01 | 辽宁万象联合医疗科技有限公司 | Quality control artificial intelligence system and quality control method for children radiation shooting and diagnosis |
CN111028939A (en) * | 2019-11-15 | 2020-04-17 | 华南理工大学 | Multigroup intelligent diagnosis system based on deep learning |
CN111028939B (en) * | 2019-11-15 | 2023-03-31 | 华南理工大学 | Multigroup intelligent diagnosis system based on deep learning |
CN111192680A (en) * | 2019-12-25 | 2020-05-22 | 山东众阳健康科技集团有限公司 | Intelligent auxiliary diagnosis method based on deep learning and collective classification |
CN111192680B (en) * | 2019-12-25 | 2021-06-01 | 山东众阳健康科技集团有限公司 | Intelligent auxiliary diagnosis method based on deep learning and collective classification |
CN111199797A (en) * | 2019-12-31 | 2020-05-26 | 中国中医科学院中医药信息研究所 | Auxiliary diagnosis model establishing and auxiliary diagnosis method and device |
CN111524570A (en) * | 2020-05-06 | 2020-08-11 | 万达信息股份有限公司 | Ultrasonic follow-up patient screening method based on machine learning |
CN111524570B (en) * | 2020-05-06 | 2024-01-16 | 万达信息股份有限公司 | Ultrasonic follow-up patient screening method based on machine learning |
CN111599462A (en) * | 2020-05-09 | 2020-08-28 | 吾征智能技术(北京)有限公司 | Intelligent body abnormal smell screening system based on cognitive learning |
CN111599462B (en) * | 2020-05-09 | 2023-11-21 | 吾征智能技术(北京)有限公司 | Intelligent body abnormal odor screening system based on cognitive learning |
CN112259229A (en) * | 2020-06-10 | 2021-01-22 | 北京小白世纪网络科技有限公司 | Method, system and equipment for realizing traditional Chinese medicine diagnosis and treatment based on neural network |
CN111883251A (en) * | 2020-07-28 | 2020-11-03 | 平安科技(深圳)有限公司 | Medical misdiagnosis detection method and device, electronic equipment and storage medium |
WO2021120688A1 (en) * | 2020-07-28 | 2021-06-24 | 平安科技(深圳)有限公司 | Medical misdiagnosis detection method and apparatus, electronic device and storage medium |
CN111739643A (en) * | 2020-08-20 | 2020-10-02 | 耀方信息技术(上海)有限公司 | Health risk assessment system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107833629A (en) | Aided diagnosis method and system based on deep learning | |
CN107016438B (en) | System based on traditional Chinese medicine syndrome differentiation artificial neural network algorithm model | |
Yang et al. | Classification of acoustic physiological signals based on deep learning neural networks with augmented features | |
CN110675944A (en) | Triage method and device, computer equipment and medium | |
CN109686441A (en) | A kind of big data medical data feature extraction and intellectual analysis prediction technique | |
CN109670179A (en) | Case history text based on iteration expansion convolutional neural networks names entity recognition method | |
CN113314205B (en) | Efficient medical image labeling and learning system | |
CN109935336A (en) | A kind of the intelligent auxiliary diagnosis method and diagnostic system of children's division of respiratory disease disease | |
CN111430025B (en) | Disease diagnosis model training method based on medical image data augmentation | |
CN110838368A (en) | Robot active inquiry method based on traditional Chinese medicine clinical knowledge graph | |
CN110427486B (en) | Body condition text classification method, device and equipment | |
KR102424085B1 (en) | Machine-assisted conversation system and medical condition inquiry device and method | |
CN110443372A (en) | A kind of transfer learning method and system based on entropy minimization | |
CN112489769A (en) | Intelligent traditional Chinese medicine diagnosis and medicine recommendation system for chronic diseases based on deep neural network | |
CN108877923A (en) | A method of the tongue fur based on deep learning generates prescriptions of traditional Chinese medicine | |
CN107785074A (en) | A kind of disease subsidiary discriminant method and system of Process Based engine | |
CN116797848A (en) | Disease positioning method and system based on medical image text alignment | |
CN114822874A (en) | Prescription efficacy classification method based on characteristic deviation alignment | |
He et al. | Differentiable automatic data augmentation by proximal update for medical image segmentation | |
CN117122303B (en) | Brain network prediction method, system, equipment and storage medium | |
CN114359656A (en) | Melanoma image identification method based on self-supervision contrast learning and storage device | |
CN107085655A (en) | The traditional Chinese medical science data processing method and system of constrained concept lattice based on attribute | |
Deng et al. | Learning‐based 3T brain MRI segmentation with guidance from 7T MRI labeling | |
CN116630964A (en) | Food image segmentation method based on discrete wavelet attention network | |
Wang et al. | SURVS: A Swin-Unet and game theory-based unsupervised segmentation method for retinal vessel |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180323 |