CN113674859A - Heart birth defect diagnosis method and system - Google Patents
Heart birth defect diagnosis method and system Download PDFInfo
- Publication number
- CN113674859A CN113674859A CN202111037705.7A CN202111037705A CN113674859A CN 113674859 A CN113674859 A CN 113674859A CN 202111037705 A CN202111037705 A CN 202111037705A CN 113674859 A CN113674859 A CN 113674859A
- Authority
- CN
- China
- Prior art keywords
- data
- medical record
- information
- heart disease
- 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
- 238000003745 diagnosis Methods 0.000 title claims abstract description 78
- 238000000034 method Methods 0.000 title claims abstract description 59
- 208000032170 Congenital Abnormalities Diseases 0.000 title claims abstract description 53
- 230000007698 birth defect Effects 0.000 title claims abstract description 53
- 206010010356 Congenital anomaly Diseases 0.000 claims abstract description 37
- 238000007689 inspection Methods 0.000 claims abstract description 19
- 238000012549 training Methods 0.000 claims abstract description 16
- 208000028831 congenital heart disease Diseases 0.000 claims description 88
- 208000002330 Congenital Heart Defects Diseases 0.000 claims description 78
- 238000004422 calculation algorithm Methods 0.000 claims description 28
- 230000002861 ventricular Effects 0.000 claims description 18
- 238000012545 processing Methods 0.000 claims description 16
- 238000007726 management method Methods 0.000 claims description 13
- 230000000747 cardiac effect Effects 0.000 claims description 10
- 230000017531 blood circulation Effects 0.000 claims description 8
- 238000013135 deep learning Methods 0.000 claims description 8
- 230000001746 atrial effect Effects 0.000 claims description 7
- 230000009466 transformation Effects 0.000 claims description 7
- 238000003066 decision tree Methods 0.000 claims description 6
- 238000001514 detection method Methods 0.000 claims description 6
- 238000005065 mining Methods 0.000 claims description 6
- 238000013523 data management Methods 0.000 claims description 5
- 210000004491 foramen ovale Anatomy 0.000 claims description 5
- 208000001910 Ventricular Heart Septal Defects Diseases 0.000 claims description 4
- 210000002376 aorta thoracic Anatomy 0.000 claims description 4
- 238000006243 chemical reaction Methods 0.000 claims description 4
- 238000012937 correction Methods 0.000 claims description 4
- 208000025339 heart septal defect Diseases 0.000 claims description 4
- 210000004115 mitral valve Anatomy 0.000 claims description 4
- 210000001147 pulmonary artery Anatomy 0.000 claims description 4
- 210000000591 tricuspid valve Anatomy 0.000 claims description 4
- 201000003130 ventricular septal defect Diseases 0.000 claims description 4
- 230000005906 menstruation Effects 0.000 claims description 3
- 208000019622 heart disease Diseases 0.000 abstract description 24
- 238000012800 visualization Methods 0.000 abstract description 4
- 230000002349 favourable effect Effects 0.000 abstract 1
- 230000006870 function Effects 0.000 description 13
- 238000010586 diagram Methods 0.000 description 10
- 238000002474 experimental method Methods 0.000 description 10
- 238000011156 evaluation Methods 0.000 description 9
- 238000004458 analytical method Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 238000011160 research Methods 0.000 description 7
- 210000002458 fetal heart Anatomy 0.000 description 6
- 238000004393 prognosis Methods 0.000 description 6
- 239000000243 solution Substances 0.000 description 6
- 230000000694 effects Effects 0.000 description 5
- 238000012384 transportation and delivery Methods 0.000 description 5
- 230000009286 beneficial effect Effects 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 3
- 238000007792 addition Methods 0.000 description 3
- 238000007405 data analysis Methods 0.000 description 3
- 238000007418 data mining Methods 0.000 description 3
- 201000010099 disease Diseases 0.000 description 3
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 3
- 210000003754 fetus Anatomy 0.000 description 3
- 238000003384 imaging method Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000009984 peri-natal effect Effects 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- 238000002604 ultrasonography Methods 0.000 description 3
- 210000000709 aorta Anatomy 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 239000003814 drug Substances 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000001605 fetal effect Effects 0.000 description 2
- 230000009067 heart development Effects 0.000 description 2
- 230000008676 import Effects 0.000 description 2
- 238000002347 injection Methods 0.000 description 2
- 239000007924 injection Substances 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000012502 risk assessment Methods 0.000 description 2
- 230000005236 sound signal Effects 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 208000024172 Cardiovascular disease Diseases 0.000 description 1
- 208000035478 Interatrial communication Diseases 0.000 description 1
- 208000006396 Pulmonary artery stenosis Diseases 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 206010002906 aortic stenosis Diseases 0.000 description 1
- 208000013914 atrial heart septal defect Diseases 0.000 description 1
- 206010003664 atrial septal defect Diseases 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000013479 data entry Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 230000002526 effect on cardiovascular system Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 210000002837 heart atrium Anatomy 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000009533 lab test Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000035935 pregnancy Effects 0.000 description 1
- 238000003793 prenatal diagnosis Methods 0.000 description 1
- 238000013138 pruning Methods 0.000 description 1
- 230000002685 pulmonary effect Effects 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
- 230000002792 vascular Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
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
- G16H50/20—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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Public Health (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Theoretical Computer Science (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Epidemiology (AREA)
- Pathology (AREA)
- Databases & Information Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention relates to a method for diagnosing birth defects of heart and a system for executing the method, which comprises the following steps: inputting data to be analyzed into a model, wherein the model is trained by a plurality of groups of training data; the training data comprises examination data and diagnosis results recorded in medical records; acquiring output information of the model, wherein the output information comprises a diagnosis result corresponding to data to be analyzed; and the establishment of the model comprises the following steps: acquiring a medical record picture; acquiring a medical record text, wherein the medical record text is acquired by identifying a medical record picture; acquiring classification information by identifying a medical record text, wherein the classification information comprises personal information and inspection information; the method has the advantages that the existing medical record data can be used for quickly constructing the prediction model, the efficient result prediction function is realized, the visualization of the heart disease prediction logic can be realized, and the method is favorable for constructing a diagnosis standard and refined assessment system of typing and layering.
Description
Technical Field
The invention relates to the technical field of medical diagnosis, in particular to a method and a system for diagnosing heart birth defects.
Background
Medicine is the most rapidly developing subject in the last two decades, with a clear tendency to specialize. Fetal heart medicine is also a 'blind area' in the medical field of China. Although obstetrics and prenatal diagnosticians can diagnose abnormal cardiac development, the obstetrics and prenatal diagnosticians do not belong to specialists in the field of cardiovascular disease of children, and do not necessarily accurately judge the subtype of abnormal cardiac development and provide the most authoritative prognostic consultation. In addition, China is wide in regions, the development of the cardiovascular field of children is uneven, and the top-level heart center of children and the heart center of common children have a 5-10 year difference in the understanding of disease diagnosis and prognosis. The outstanding clinical problem exists at present is that diagnosis and clinical evaluation of the fetal heart birth defects lack a diagnosis specification and a detailed evaluation system of parting and layering, so that clear diagnosis and treatment consultation guidance can not be provided for prenatal diagnosticians and infant family members, and proper transportation and multidisciplinary integrated management of the perinatal period of the fetal heart birth defects are not facilitated.
The existing research is mainly aimed at the image research and the heart sound signal research of the congenital heart disease, because the existing research does not have the support of a large amount of case data of the congenital heart disease and the complexity of a case data structure, the evaluation of the congenital heart disease does not have systematicness and standardization, and meanwhile, the knowledge mining cannot be realized from a large amount of clinical data. The existing classification algorithm and data mining algorithm for other diseases cannot have direct excellent performance due to the uniqueness of the data of the medical record of the congenital heart disease.
At present, the diagnosis of the congenital heart disease is mainly divided into direct diagnosis of characteristic data of the heart disease, diagnosis on an image for ultrasonic detection of a fetus and analysis of heart sound data, such as a heart disease auxiliary diagnosis system based on a k-neighbor algorithm to improve the accuracy of heart disease diagnosis and improve the real-time performance. The change of the cardiac axis is measured by using an echocardiogram, and a doctor directly analyzes some valuable information of clinical data through manual work on an ultrasonic image and uses the information for a usual prenatal diagnosis basis or research on the ultrasonic image, such as auxiliary diagnosis of fetal heart defects by using a real-time dynamic four-dimensional time-space correlation imaging technology. The heart defect problem of the fetus is diagnosed by utilizing a nine-section analysis method and the application of a double-low scanning technology to the vascular imaging of the congenital heart disease of the children is utilized.
Therefore, the current auxiliary diagnosis about congenital heart disease has the problem that the direct auxiliary diagnosis means for the fetus with congenital heart defects is relatively limited, and most doctors manually find the existence rules of the congenital heart disease according to the characteristic data. At present, the heart disease is diagnosed only by means of the image display of the ultrasound and the real-time prediction of the classification condition of the heart disease and the analysis of the heart sound signals through a deep learning method. Various semantic information is extracted from clinical medical records, inspection records and laboratory test reports stored for the congenital heart disease, so that the problem that the current research is not comprehensive enough in application scenes such as medical scientific research analysis and the like is solved. The characterization of the heart disease is less consistent with the results presented by different physicians in different hospital settings. There is currently no uniform and definite assessment criteria for the consensus of a definite congenital heart disease that has been used, and that can be professionally scored against the results of a definite diagnosis of a congenital heart disease given by a physician, so that there are clear standard and standardized guidelines for the physician and the parents of the congenital heart disease child and standardized administration recommendations for providing pregnancy.
The Chinese patent application: CN103116707A discloses an intelligent diagnosis method for heart disease based on case reasoning, which searches various examination indexes of a patient as a case in a case library, finds out the most similar record as a diagnosis result, obtains a distribution scheme of each characteristic attribute weight of the case by using a water injection principle, and eliminates redundant attributes according to the attribute weights. Thereby improving the accuracy and speed of the diagnosis result of the heart disease, comprising: defining a case representation; constructing a historical case set; carrying out weight distribution on the characteristic attributes by using a water injection principle; extracting the characteristic attribute through the distribution result of the attribute weight; calculating case similarity; reusing the matched case; correcting the case; the patent application improves the retrieval strategy in case inference, and ensures the diagnosis precision and speed by optimizing and distributing the case characteristic attribute weight and carrying out attribute reduction link; according to the method, a standard case library is essentially constructed, diagnosis results are analogized through the similarity between cases, and are provided for medical staff to refer, although the program is simple, the determinant factor of the congenital heart disease cannot be found, the prediction result is indirect, the efficiency is low, and the judgment basis of the congenital heart disease cannot be displayed for the medical staff in a large medical record.
In summary, there is a need for a method and a system for diagnosing cardiac birth defects, which can utilize the existing medical record data to quickly construct a prediction model, have an efficient result prediction function, can realize the visualization of the prediction logic of congenital heart disease, and are helpful for constructing a typing and layering diagnosis specification and a detailed assessment system.
Disclosure of Invention
The invention aims to provide a heart birth defect diagnosis method which can utilize the existing medical record data to quickly construct a prediction model, has an efficient result prediction function, can realize the visualization of the heart disease prediction logic and is beneficial to constructing a parting and layering diagnosis specification and refining evaluation system.
In order to achieve the purpose, the invention adopts the technical scheme that:
a method for diagnosing birth defects of the heart, comprising: inputting data to be analyzed into a model, wherein the model is trained by using a plurality of groups of training data; each of the plurality of sets of training data includes examination data and diagnostic results recorded in a medical record; and acquiring output information of the model, wherein the output information comprises a diagnosis result corresponding to the data to be analyzed.
As a preferred technical solution, the establishing of the model includes: acquiring a medical record picture; acquiring a medical record text, wherein the medical record text is acquired by identifying the medical record picture; acquiring classification information, wherein the classification information is obtained by identifying the medical record text, and the classification information comprises personal information and inspection information; the examination information comprises the examination data and diagnosis results; and establishing a model, wherein the model is obtained by deep learning of the inspection information.
As a preferred technical solution, the method for diagnosing the birth defects of the heart further comprises outputting a heart disease judgment logic.
As a preferred technical solution, the outputting the congenital heart disease judgment logic includes: outputting classification information, wherein the classification information comprises the personal information and examination information, and the examination information consists of common examination information and special examination information; outputting a congenital heart disease judgment rule, wherein the congenital heart disease rule is obtained based on a C4.5 decision tree; outputting a deterministic characteristic, wherein the deterministic characteristic is obtained based on an association rule algorithm.
As a preferred technical scheme, the medical record pictures comprise paper medical record pictures and electronic medical record pictures, and the original data format is converted into a jpg format.
As a preferred technical solution, the acquiring a medical history text includes: carrying out correction transformation through Hough line detection and perspective transformation; and performing character recognition based on Tesseract-OCR.
As a preferred technical solution, the personal information data includes but is not limited to age, last menstruation, date of examination; the common examination information includes, but is not limited to, mandrel, cardiothoracic ratio, heart rate, left atrial transverse diameter, right atrial transverse diameter, aortic inner diameter, pulmonary artery inner diameter, left ventricular transverse diameter to right ventricular transverse diameter, left ventricular transverse diameter, right ventricular transverse diameter; the unique inspection information includes, but is not limited to: foramen ovale inter-radial diameter, mitral valve blood flow, tricuspid valve blood flow, primary septal defect, ventricular septal defect, aortic arch inner diameter, catheter arch inner diameter, ventricular septal echo interruption.
As a preferable technical scheme, the establishing model is obtained based on an XGboost algorithm.
Another objective of the present invention is to provide a system for diagnosing birth defects of the heart, which can utilize the existing medical record data to quickly construct a prediction model, has a highly efficient result prediction function, can realize the visualization of the prediction logic of the congenital heart disease, and is helpful for constructing a parting and layering diagnosis specification and a detailed evaluation system.
In order to achieve the purpose, the invention adopts the technical scheme that:
a system for diagnosing birth defects of the heart, comprising: the device comprises an input module, an output module and a data processing module; the input module is used for inputting information, including input of data to be analyzed and medical record pictures; the output module is used for outputting information, and comprises a diagnosis result and output of a congenital heart disease judgment logic, wherein the diagnosis result further comprises a diagnosis suggestion; the congenital heart disease judgment logic comprises classification information, a congenital heart disease judgment rule and a decisive characteristic; the data processing module is used for establishing a model and processing data; the system for diagnosing the cardiac birth defect performs the method for diagnosing the cardiac birth defect according to any one of claims 1-8.
As a preferred technical solution, the input module includes: a medical record data management module and a system user personal information management module; the output module includes: the heart defect assessment and suggestion module and the congenital heart disease judgment logic display module are connected with the heart defect assessment and suggestion module; the data processing module comprises: the system comprises a medical record identification and conversion module, a classification auxiliary diagnosis module of the congenital heart disease and a congenital heart disease judgment logic mining module.
The invention has the advantages that:
the heart birth defect diagnosis method and the system of the invention obtain the heart disease prediction model based on the medical record sample database, can provide the heart disease prediction result for the reference of medical care personnel, the sample database is constructed in the form of medical record pictures, the input construction burden of the sample database is greatly reduced, and simultaneously, a huge number of paper medical records are utilized, so that the finally obtained prediction model is more accurate and effective; the method and the system for diagnosing the birth defects of the heart can also output classification information, the judgment rule of the congenital heart disease and decisive characteristics, visualize the judgment logic of the congenital heart disease formed based on big data, are beneficial to constructing a diagnosis standard and detailed evaluation system of typing and layering, simultaneously improve the reliability of a diagnosis result given by medical personnel, and further optimize a medical record sample database, so that the existing model is corrected to have more accurate prediction capability, provide more accurate association rules according to corresponding algorithms, realize more reliable virtuous cycle of the diagnosis result, and be beneficial to providing safer and more efficient prognosis consultation for patients.
Drawings
Fig. 1 is a flow chart of a method for diagnosing birth defects of a heart in accordance with the present invention.
Fig. 2 is a flow chart of the model building according to the method for diagnosing the birth defects of the heart of the invention.
Fig. 3 is a flow chart of another method for diagnosing birth defects of a heart in accordance with the present invention.
Fig. 4 is a logic diagram for obtaining the precedents judgment rule of another method for diagnosing the birth defects of the heart according to the present invention.
Fig. 5 is a block diagram of a system for diagnosing birth defects of the heart in accordance with the present invention.
Fig. 6 is a frame diagram of an application of the system for diagnosing birth defects of heart of the present invention.
FIG. 7 is a schematic diagram of a user information management interface of the embodiment.
FIG. 8 is a schematic diagram of a medical record information management interface according to this embodiment.
FIG. 9 is a schematic view of the prognostic diagnostic interface of this embodiment.
FIG. 10 is a schematic diagram of an evaluation and recommendation interface according to this embodiment.
Detailed Description
The invention will be further illustrated with reference to specific embodiments. It should be understood that these examples are for illustrative purposes only and are not intended to limit the scope of the present invention. Furthermore, it should be understood that various changes and modifications can be made by those skilled in the art after reading the disclosure of the present invention, and equivalents fall within the scope of the appended claims.
The reference numerals and components referred to in the drawings are as follows:
01. input module 02, output module 03 data processing module
Example 1
Referring to fig. 1, fig. 1 is a flow chart of a method for diagnosing birth defects of heart according to the present invention.
A method for diagnosing birth defects of heart at least comprises the following steps S10-S20:
step S10: inputting data to be analyzed into a model, wherein the model is trained by using a plurality of groups of training data; each of the plurality of sets of training data includes examination data and diagnostic results recorded in a medical record;
the data to be analyzed is physical examination data of a patient or a physical examination person, the model is used for generating a corresponding diagnosis result according to the input data to be analyzed, the model is obtained by a deep learning algorithm through a large amount of data, and the adopted training data is records of existing related medical records, and comprises a large amount of antecedent disease medical record data and a large amount of normal medical record data; the deep learning algorithm generates a congenital heart disease prediction model by learning the corresponding relation between a plurality of examination data in a medical record and a diagnosis result thereof.
Step S20: and acquiring output information of the model, wherein the output information comprises a diagnosis result corresponding to the data to be analyzed.
When the data to be analyzed is input into the model, a predicted diagnosis result corresponding to the data to be analyzed can be obtained, and the diagnosis result can be the presence of the congenital heart disease or the absence of the congenital heart disease; even outputting specific suspected congenital heart disease types, such as pulmonary artery stenosis, aortic stenosis, atrial septal defect and the like; or the probability of suspected congenital heart disease can be output, for example, the probability of suffering from congenital heart disease is 80%, 10% and the like; preferably, the model can also output corresponding diagnosis suggestions, such as delivery suggestions related to pregnant and lying-in women; the specific output mode depends on the application of a deep learning algorithm and the expansion of a training data sample base, so that the model is more accurate.
Referring to fig. 2, fig. 2 is a flow chart of modeling for a method for diagnosing birth defects of heart according to the present invention. In the present embodiment, the establishment of the prediction model of the heart disease at least includes the following steps S100 to S106:
step S100: acquiring a medical record picture;
in order to facilitate the establishment of the training data sample library, the medical record is subjected to imaging processing, wherein the medical record images comprise paper medical record images and electronic medical record images; and the corresponding pictures are all converted into pictures in a jpg format, for example, pictures in various formats, pdf documents and the like are uniformly converted into the jpg format for processing.
Step S102: acquiring a medical record text, wherein the medical record text is acquired by identifying the medical record picture;
preferably, the acquiring the medical record text comprises the following steps S1020 to S1022:
step S1020: carrying out correction transformation through Hough line detection and perspective transformation;
due to the fact that the situation that paper medical records are placed improperly in the data acquisition process, namely the scanning process, the obtained medical record picture data are seriously inclined, and the accuracy of character recognition in the later period is low; the embodiment solves the problem of inclination of medical record pictures through Hough line detection and perspective transformation correction.
Step S1022: performing character recognition based on Tesseract-OCR;
specifically, the method comprises the following steps: firstly, analyzing the page layout of a text region of the obtained binary image to find out a corresponding text character region and a corresponding sub-region, and forming a required block region by all corresponding contour sets; detecting the outline of the character through the block area obtained previously so as to obtain a required text line and obtain a corresponding text line and word area through a space area; thirdly, using a self-adaptive classifier to perform word analysis on the obtained text line and word area twice; and fourthly, because fuzzy spaces exist in the word area, the recognition accuracy is improved by the modes of fuzzy spaces, stroke height detection, lower case letters and the like, and finally final text data, namely the acquired case text, is obtained.
It should be noted that: the most difficult part in the implementation of the whole architecture is the segmentation and recognition work of the characters. The character segmentation process obtains the size of a text and the area where the text is located through layout analysis of a page, if the cross-row condition exists, the cross-row self-drainage middle value height is obtained through a percentile height filter, and meanwhile, some noise areas such as punctuation marks and the like are removed. And then, fitting a straight line to the obtained block area, mainly obtaining a more accurate text line shape by selecting the variance with the minimum median, and then improving the fitting accuracy by a least square method. At the same time, the areas of the text lines that are stuck to each other need to be segmented. The character is divided mainly by using the existing blank area between the characters and the character characteristics of equal intervals. The other complicated work is to identify the character region obtained by segmentation, firstly, the topological feature, the approximate polygon feature and the contour feature syntactic feature of the character need to be extracted, and the comparison is carried out with the training library prototype in a many-to-one mode. The whole process is divided into two processes of rough division and subdivision, wherein the rough division is to find out characters with similar characteristics, and then subdivide through calculating the distance between the characteristics. Due to the adoption of the adaptive classifier, the recognition efficiency of the whole page is higher than that of a single character in the whole implementation process.
Step S104: acquiring classification information, wherein the classification information is obtained by identifying the medical record text, and the classification information comprises personal information and inspection information; the examination information comprises the examination data and diagnosis results;
the method extracts the required features from the acquired medical record text data, so that subsequent classification prediction and knowledge discovery are facilitated, namely, a corresponding model is established; and because the medical record data of the congenital heart disease is relatively normalized, the related information can be acquired by adopting a regular matching mode: personal information and inspection information; the personal information of the patient is mainly obtained by regular matching, such as the age of the patient, the last menstruation of the patient, the inspection date of the patient, whether the ultrasound is obviously abnormal or not and the like; the examination information of the patient is mainly obtained by regular matching, such as fetal position, mandrel, cardiothoracic ratio, heart rate, left atrial transverse diameter, right atrial transverse diameter, aortic internal diameter, pulmonary internal diameter, left ventricular transverse diameter ratio to right ventricular transverse diameter, left ventricular transverse diameter, right ventricular transverse diameter, oval foramen intervallic diameter, mitral valve blood flow, tricuspid valve blood flow, primary septal defect, ventricular septal defect, aortic arch internal diameter, catheter arch internal diameter, ventricular septal echo interruption and the like; the diagnosis result is the corresponding final diagnosis result and diagnosis suggestion recorded by each medical record.
Step S106: establishing a model, wherein the model is obtained by deep learning of the inspection information;
in the embodiment, an XGboost algorithm is adopted to carry out learning processing on the inspection information; specifically, the following algorithm steps are executed:
1. the spare _ rate and tree _ based parameters are determined. In order to determine the Boosting parameter, the experiment is performed with initial values. Firstly, values are taken according to the following method:
(a) the maximum depth of the tree is set to 5.
(b) The weight experiment for the minimal node is initialized to 1: the prediction of the heart disease chooses a smaller value because the prediction of the heart disease is a classification problem of extreme imbalance. The values under some leaf nodes will be smaller.
(c) The minimum loss function drop value is set to 0, and the smaller value of the initial value setting is achieved because this parameter is subsequently or needs to be adjusted.
(d) The ratio of the number of samples per tree to the number of randomly sampled columns per tree is set to 0.5.
(e) scale _ pos _ weight is set to 1, which allows for fast convergence.
(f) Here, the learning rate is set to 0.1. The experiment was followed by working with the CV function to obtain the optimal number of trees required.
2. The maximum tree depth and the minimum weight parameter are adjusted. The experiment optimizes the two parameters firstly because the experiment can firstly estimate the parameters with coarse granularity and then adjust the parameters with fine granularity, so that the two factors which have the greatest influence on the experimental result are optimal. Experiments here would perform a high load grid search, mainly by defining the ranking as 'roc _ auc' because it supports two classes.
3. And optimizing the parameter of the reduction value of the minimum loss function. After the maximum tree depth and the minimum weight parameters are determined, the experiment starts to adjust the minimum loss function degradation value parameters. The minimum weight parameter has a wide range of values, all set to 5 in the experimental procedure here.
4. And adjusting the ratio of the number of the sub-samples of each tree and the column number parameter of each random sample. Different sub-sample per tree ratios and column number per random sample parameters were tried. We proceed with this step in two stages. Both steps take 0.6,0.7,0.8,0.9 as starting values.
5. And adjusting and optimizing the regularization parameters. Here we tune the gamma function as it provides a more efficient way to reduce overfitting.
6. The learning rate is reduced. The experiment was followed by working with the CV function to obtain the optimal number of trees required.
The accuracy of classification prediction is 85.16%, the accuracy is 81.82% and the recall rate is 83.33%.
It should be noted that: there are many integrated algorithms for predictive classification, Boosting, AdaBoost, GBDT algorithms and XGBoost algorithms. Boosting is the integration algorithm that was proposed relatively early, and it mainly achieves the effect of a strong classifier after setting the corresponding weights by integrating many weak classifiers. The whole implementation process is that the classifiers are added into the classifier one by one, and then the addition is stopped after the best prediction effect is achieved. AdaBoost is a model implementation used mainly for binary problems. The algorithm can break the original sample distribution, and an optimal classifier is searched by a backtracking method, so that the classification effect is optimal. The GBDT algorithm is realized by constructing a plurality of trees, then comparing the obtained predicted values and actual values of the trees, then calculating residual values, and finally, continuously iterating the learned trees to enable the prediction effect of the trees to be optimal. The XGboost algorithm is characterized in that after a plurality of trees are gathered together, a CART regression tree model is changed into a strong classifier, the XGboost algorithm is more excellent than a GBDT algorithm in that the XGboost algorithm applies regularization terms and other operations to prevent overfitting, a second derivative is used for optimizing a loss function so as to optimize an objective function, the situation that data is excessive in default is well processed, the training speed can be well controlled, and when a good tree is established, the number of constructed trees is stopped. Therefore, in the classification prediction experiment of the congenital heart disease, the XGBoost algorithm is taken as an example to realize the establishment of the model in the embodiment.
According to the method for diagnosing the cardiac birth defect, the early heart disease prediction model is constructed through deep learning of the early heart disease medical record data, the prediction result is provided for medical care personnel to refer to, the medical record data is picturized, the utilization rate of resources is greatly improved, the difficulty in manually establishing a huge sample database is reduced, a data basis is provided for establishing an efficient early heart disease prediction model, the final diagnosis result tends to be more accurate, and more targeted prognosis consultation is facilitated.
Example 2
Referring to fig. 3, fig. 3 is a flow chart of another method for diagnosing birth defects of a heart according to the present invention. In order to enable medical staff to obtain the congenital heart disease rule based on the sample database, the diagnosis conditions of the congenital heart disease can be better known from the data;
preferably, the method for diagnosing the birth defects of the heart further comprises step S14:
step S14: and outputting the classification information, wherein the classification information comprises the personal information and the inspection information, and the inspection information comprises common inspection information and specific inspection information.
The examination information includes all examination information recorded in a medical record, wherein not all data information is related to the congenital heart disease, and the examination information can be divided into common examination information and specific examination information according to analysis of a large amount of data, wherein the common examination information is diagnosis data unrelated to the prognosis of the congenital heart disease, such as fetal orientation, mandrel, cardiothoracic ratio, heart rate, left transverse diameter, right transverse diameter of the atrium, internal diameter of aorta, internal diameter of pulmonary artery, left transverse diameter of the ventricle to right transverse diameter of the ventricle, left transverse diameter of the ventricle, right transverse diameter of the ventricle and the like; the specific examination information is diagnostic data related to the prediction of the congenital heart disease, such as the foramen ovale diameter, mitral valve blood flow, tricuspid valve blood flow, primary septal defect, ventricular septal defect, aortic arch inner diameter, catheter arch inner diameter, ventricular septal echo interruption and the like.
Preferably, the method for diagnosing the birth defects of the heart further comprises step S16:
step S16: outputting a congenital heart disease judgment rule, wherein the congenital heart disease rule is obtained based on a C4.5 decision tree;
the precedental heart disease judgment rule is a precedental heart disease diagnosis logic which is acquired based on a large amount of medical record sample data, the obtained judgment logic can be a plurality of items if each piece of inspection information comprises an inspection item name and a corresponding judgment threshold, and the items can be displayed.
In the present embodiment, please refer to fig. 4, wherein fig. 4 is a logic diagram illustrating a determination rule of congenital heart disease in another method for diagnosing birth defects of a heart according to the present invention. In order to better obtain the classification rule of the congenital heart disease, the congenital heart disease judgment rule is obtained based on a C4.5 decision tree, negative medical record data with the attribute default number exceeding 3 are directly discarded, the remaining data are finally used for training a model, and in order to prevent the over-fitting condition of the decision tree result and obtain a stronger rule, so that the classification accuracy is discarded, the maximum depth of the tree is set to be 4, and the minimum number of samples required by internal node subdivision is 8.
Preferably, the method for diagnosing the birth defects of the heart further comprises step S18:
step S18: outputting a deterministic characteristic, wherein the deterministic characteristic is obtained based on an association rule algorithm.
The decisive characteristic data is examination information with large correlation to the congenital heart disease, and the method comprises the following steps: firstly, finding all frequent item sets, namely finding all data with equal attribute values corresponding to the features in the whole data set of the congenital heart disease, then counting, and selecting items with the support degree greater than the minimum support degree. Naturally, the two item sets most frequently in the heart disease are followed by the corresponding two item sets with all 1 and all 0 labeled attributes, but all the item sets must be found to be more than the minimum support. And secondly, generating a candidate set by combining the found frequent item sets, so that the condition that the operation time is increased due to the fact that an infrequent set is found is avoided, the pruning effect is achieved, and all transactions are scanned to count all items in the candidate set. And finding out all item sets meeting the minimum support degree and the minimum credibility from the combined candidate sets. And thirdly, repeating the operation of the second step from the item set selected in the second step so as to find all the item sets meeting the minimum support degree and the minimum confidence degree. Thereby finding a strong association rule.
For example, from the rules found in the decision tree, the present invention finds that there are different left atrial transverse diameter, right transverse diameter, left ventricular transverse diameter, right ventricular transverse diameter, foramen ovale diameter, pulmonary artery inner diameter, aorta inner diameter, etc. at different gestational week times, which can search for the range value of the fetal heart normal and the determining characteristics of the fetal heart defect under different gestational week conditions.
According to the heart birth defect diagnosis method, through the output of the classification information, the congenital heart disease judgment rule and the decisive characteristics, the diagnosis rule and logic of the congenital heart disease obtained based on medical record sample data are presented in a visual form, and medical workers can compare and study the diagnosis result by combining own experience, so that the final diagnosis and prediction tend to be more accurate, a more targeted treatment scheme can be formed, the more accurate prediction result can be used as new sample data, the model is further corrected, and a typing and layering diagnosis standard and refined evaluation system is favorably established.
Example 3
Referring to fig. 5, fig. 5 is a block diagram of a system for diagnosing birth defects of heart according to the present invention. A heart birth defect diagnosis system for performing the heart birth defect diagnosis method, comprising: the device comprises an input module 01, an output module 02 and a data processing module 03;
the input module 01 is used for inputting information, including input of data to be analyzed and medical record pictures; the output module 02 is used for outputting information, including a diagnosis result and output of a congenital heart disease judgment logic, wherein the diagnosis result further includes a diagnosis suggestion; the congenital heart disease judgment logic comprises classification information, a congenital heart disease judgment rule and a decisive characteristic; the data processing module 03 is used for establishing a model and processing data;
referring to fig. 6, fig. 6 is a frame diagram of a system for diagnosing birth defects of heart according to the present invention. Specifically, the application mode comprises a data layer, a data analysis layer, an application layer and a presentation layer, wherein the data layer is used for storing related data such as medical record pictures, medical records in other forms and algorithm logic; the data analysis layer is used for calling related algorithm logic to analyze the medical record picture, constructing and updating the model and calculating related results; the application layer comprises algorithm management, database management and user information management and is used for updating related information; the presentation layer comprises a WEB front end used for displaying relevant results, and the input module can import corresponding data through the application layer, wherein the import module comprises but is not limited to the input of data to be analyzed and the input of medical record pictures, and can also comprise the updating of algorithm logic and the like.
The heart birth defect diagnosis system can execute the heart birth defect diagnosis method, so that a user can visually acquire corresponding information; in this embodiment, an application mode of the method and system for diagnosing the cardiac birth defect is provided through the construction of a data layer, a data analysis layer, an application layer and a presentation layer, and the method and system are simple in structure and beneficial to implementation.
Example 4
In this embodiment, the system and the method for diagnosing the cardiac birth defect are described from a client application. It should be noted that the present embodiment only illustrates the basic operation mode, and all applications of the method and system for diagnosing cardiac birth defects of the present invention are within the scope of the present invention.
The input module of the system for diagnosing birth defects of heart in this embodiment comprises: a medical record data management module and a system user personal information management module; its output module includes: the heart defect assessment and suggestion module and the congenital heart disease judgment logic display module are connected with the heart defect assessment and suggestion module; the data processing module comprises: the system comprises a medical record identification and conversion module, a classification auxiliary diagnosis module for congenital heart disease and a congenital heart disease judgment logic mining module;
referring to fig. 7, fig. 7 is a schematic view of a user information management interface according to the embodiment; the system user personal information management module can set different use authorities for different users and manage the management of some basic information, such as names, telephones, account numbers, passwords and the like of individuals; as shown in fig. 7: the whole system comprises three identities of system maintainers, medical staff and patients, and for different identities, the background corresponds to different system use authorities;
the medical record data management module can manage data with unified corresponding formats, upload of the data and query of corresponding data can be realized through the medical record data management module, and a normalized data access and input change are provided; as shown in fig. 8, fig. 8 is a schematic view of a medical record information management interface according to the embodiment; the module can be connected with a database to directly manage the database on a webpage, such as adding, deleting, modifying, searching and other work of medical record information;
the medical record identification and conversion module comprises: mainly aiming at the treatment of paper medical records which are not completely normalized, the module can convert the paper image medical records into text medical records, namely, the case texts are obtained through case pictures, and then the case texts are unified with the normalized training data samples through the extraction and the correspondence of classified information, so that the subsequent treatment is facilitated;
the classification auxiliary diagnosis module for the congenital heart disease is mainly used for training examination information of medical records, acquiring a congenital heart disease judgment model, acquiring data to be analyzed input by an input module, performing operation through the model, and outputting a diagnosis result through an output module, so that a doctor is assisted to provide defect problem prediction of the congenital heart disease for each new patient; FIG. 9 is a schematic view of a classification and prediction diagnostic interface according to the present embodiment;
the congenital heart disease judgment logic mining module is mainly used for mining medical record texts stored in a medical record sample database to find clinically valuable information and knowledge hidden behind the module, such as a congenital heart disease judgment rule, a decisive characteristic and the like;
referring to fig. 10, fig. 10 is a schematic diagram of an evaluation and suggestion interface of the embodiment; the heart defect assessment and suggestion module is mainly used for giving corresponding assessment scores, risk indexes (including perinatal risk assessment and long-term prognosis assessment), delivery suggestions and delivery room suggestions to patients with heart defects according to the results (obtained by diagnosis results, early heart disease judgment logic, early heart disease guide and diagnosis and treatment experience provided by a doctor comprehensive model) diagnosed by doctors;
the congenital heart disease judgment logic display module can display the congenital heart disease judgment rules, the decisive characteristics and the like acquired according to the medical record sample database, and it needs to be explained that the congenital heart disease judgment logic display module belongs to the viewing permission of medical staff.
It should be noted that the system is implemented by using a B/S architecture, which makes the user more convenient and faster to use, without any software and high performance requirements on the system. A user using the system only needs to log in a webpage and then perform basic operations such as data entry, addition, deletion, modification, check and the like on the webpage, and can also perform classified prediction of congenital heart diseases on the webpage and give corresponding assessment scores, risks, delivery suggestions, delivery room suggestions and the like of the heart birth defects according to a diagnosis result. And the importing of new medical record data can update data mining to discover more temporary production knowledge. The user can see the corresponding ultrasound image, etc. The above are all the main functions of the system, and the most key functions are classified diagnosis and prediction of the congenital heart disease, data mining of the congenital heart disease, defect assessment of the congenital heart disease after diagnosis, perinatal risk assessment, long-term prognosis assessment and relevant medical advice giving.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and additions can be made without departing from the principle of the present invention, and these should also be considered as the protection scope of the present invention.
Claims (10)
1. A method for diagnosing birth defects of the heart, comprising:
inputting data to be analyzed into a model, wherein the model is trained by using a plurality of groups of training data; each of the plurality of sets of training data includes examination data and diagnostic results recorded in a medical record;
and acquiring output information of the model, wherein the output information comprises a diagnosis result corresponding to the data to be analyzed.
2. The method for diagnosing birth defects of the heart according to claim 1, wherein the establishing of the model comprises:
acquiring a medical record picture;
acquiring a medical record text, wherein the medical record text is acquired by identifying the medical record picture;
acquiring classification information, wherein the classification information is obtained by identifying the medical record text, and the classification information comprises personal information and inspection information; the examination information comprises the examination data and diagnosis results;
and establishing a model, wherein the model is obtained by deep learning of the inspection information.
3. The method for diagnosing heart birth defects according to claim 2, further comprising outputting a congenital heart disease judgment logic.
4. The method for diagnosing birth defects of heart according to claim 3, wherein outputting the congenital heart disease judgment logic comprises:
outputting classification information, wherein the classification information comprises the personal information and examination information, and the examination information consists of common examination information and special examination information;
outputting a congenital heart disease judgment rule, wherein the congenital heart disease rule is obtained based on a C4.5 decision tree;
outputting a deterministic characteristic, wherein the deterministic characteristic is obtained based on an association rule algorithm.
5. The method as claimed in claim 2, wherein the medical record pictures include paper medical record pictures and electronic medical record pictures, and the original data format is converted into jpg format.
6. The method for diagnosing birth defects of the heart according to claim 2, wherein the acquiring of medical history text comprises:
carrying out correction transformation through Hough line detection and perspective transformation;
and performing character recognition based on Tesseract-OCR.
7. The method for diagnosing birth defects of the heart according to claim 4, wherein said personal information data includes, but is not limited to, age, last menstruation, date of examination; the common examination information includes, but is not limited to, mandrel, cardiothoracic ratio, heart rate, left atrial transverse diameter, right atrial transverse diameter, aortic inner diameter, pulmonary artery inner diameter, left ventricular transverse diameter to right ventricular transverse diameter, left ventricular transverse diameter, right ventricular transverse diameter; the unique inspection information includes, but is not limited to: foramen ovale inter-radial diameter, mitral valve blood flow, tricuspid valve blood flow, primary septal defect, ventricular septal defect, aortic arch inner diameter, catheter arch inner diameter, ventricular septal echo interruption.
8. A method for diagnosing birth defects of the heart according to claim 2, wherein said established model is obtained based on the XGBoost algorithm.
9. A system for diagnosing birth defects of the heart, comprising: the device comprises an input module, an output module and a data processing module;
the input module is used for inputting information, including input of data to be analyzed and medical record pictures;
the output module is used for outputting information, and comprises a diagnosis result and output of a congenital heart disease judgment logic, wherein the diagnosis result further comprises a diagnosis suggestion; the congenital heart disease judgment logic comprises classification information, a congenital heart disease judgment rule and a decisive characteristic;
the data processing module is used for establishing a model and processing data;
the system for diagnosing the cardiac birth defect performs the method for diagnosing the cardiac birth defect according to any one of claims 1-8.
10. The system of claim 9, wherein the input module comprises: a medical record data management module and a system user personal information management module;
the output module includes: the heart defect assessment and suggestion module and the congenital heart disease judgment logic display module are connected with the heart defect assessment and suggestion module;
the data processing module comprises: the system comprises a medical record identification and conversion module, a classification auxiliary diagnosis module of the congenital heart disease and a congenital heart disease judgment logic mining module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111037705.7A CN113674859A (en) | 2021-09-06 | 2021-09-06 | Heart birth defect diagnosis method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111037705.7A CN113674859A (en) | 2021-09-06 | 2021-09-06 | Heart birth defect diagnosis method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113674859A true CN113674859A (en) | 2021-11-19 |
Family
ID=78548702
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111037705.7A Pending CN113674859A (en) | 2021-09-06 | 2021-09-06 | Heart birth defect diagnosis method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113674859A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116189905A (en) * | 2023-02-28 | 2023-05-30 | 中国医学科学院阜外医院 | Aortic constriction prenatal diagnosis model building method and diagnosis system |
| CN116230243A (en) * | 2023-05-04 | 2023-06-06 | 四川大学华西医院 | Plugging operation diagnosis and treatment system for refractory epileptic patient with high shunt capacity PFO patient |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050020903A1 (en) * | 2003-06-25 | 2005-01-27 | Sriram Krishnan | Systems and methods for automated diagnosis and decision support for heart related diseases and conditions |
| CN107785075A (en) * | 2017-11-01 | 2018-03-09 | 杭州依图医疗技术有限公司 | Fever in children disease deep learning assistant diagnosis system based on text case history |
| CN111260209A (en) * | 2020-01-14 | 2020-06-09 | 山东大学 | Cardiovascular disease risk prediction and evaluation system combining electronic medical record and medical image |
-
2021
- 2021-09-06 CN CN202111037705.7A patent/CN113674859A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050020903A1 (en) * | 2003-06-25 | 2005-01-27 | Sriram Krishnan | Systems and methods for automated diagnosis and decision support for heart related diseases and conditions |
| CN107785075A (en) * | 2017-11-01 | 2018-03-09 | 杭州依图医疗技术有限公司 | Fever in children disease deep learning assistant diagnosis system based on text case history |
| CN111260209A (en) * | 2020-01-14 | 2020-06-09 | 山东大学 | Cardiovascular disease risk prediction and evaluation system combining electronic medical record and medical image |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116189905A (en) * | 2023-02-28 | 2023-05-30 | 中国医学科学院阜外医院 | Aortic constriction prenatal diagnosis model building method and diagnosis system |
| CN116230243A (en) * | 2023-05-04 | 2023-06-06 | 四川大学华西医院 | Plugging operation diagnosis and treatment system for refractory epileptic patient with high shunt capacity PFO patient |
| CN116230243B (en) * | 2023-05-04 | 2023-07-21 | 四川大学华西医院 | Plugging operation diagnosis and treatment system for refractory epileptic patient with high shunt capacity PFO patient |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10929420B2 (en) | Structured report data from a medical text report | |
| Zhang et al. | The comparison of LightGBM and XGBoost coupling factor analysis and prediagnosis of acute liver failure | |
| US20190191988A1 (en) | Screening method for automated detection of vision-degenerative diseases from color fundus images | |
| WO2021227511A1 (en) | Complication onset risk prediction method and system based on electronic medical record big data, and terminal and storage medium | |
| RU2686627C1 (en) | Automatic development of a longitudinal indicator-oriented area for viewing patient's parameters | |
| WO2023078025A1 (en) | Task decomposition strategy-based auxiliary differential diagnosis system for fever of unknown origin | |
| CN111292821A (en) | Medical diagnosis and treatment system | |
| US20090080732A1 (en) | System and Method for Performing Probabilistic Classification and Decision Support Using Multidimensional Medical Image Databases | |
| Ribeiro et al. | An association rule-based method to support medical image diagnosis with efficiency | |
| CN113674859A (en) | Heart birth defect diagnosis method and system | |
| CN110277167A (en) | The Chronic Non-Communicable Diseases Risk Forecast System of knowledge based map | |
| JP2014029644A (en) | Similar case retrieval device and similar case retrieval method | |
| CN111080643A (en) | Method and device for classifying diabetes and related diseases based on fundus images | |
| CN112434718B (en) | New coronary pneumonia multi-modal feature extraction fusion method and system based on depth map | |
| CN112466462B (en) | EMR information association and evolution method based on deep learning of image | |
| CN115578783B (en) | Device and method for identifying eye diseases based on eye images and related products | |
| WO2021008601A1 (en) | Method for testing medical data | |
| Chen et al. | An causal XAI diagnostic model for breast cancer based on mammography reports | |
| Kumar et al. | Lungcov: A diagnostic framework using machine learning and Imaging Modality | |
| CN113408603B (en) | Coronary artery stenosis degree identification method based on multi-classifier fusion | |
| CN114191665A (en) | Method and device for classifying man-machine asynchronous phenomena in mechanical ventilation process | |
| Zebari et al. | Automated Detection of Covid-19 from X-ray Using SVM | |
| CN113254609A (en) | Question-answering model integration method based on negative sample diversity | |
| Cao et al. | The research on medical image classification algorithm based on PLSA-BOW model | |
| Peng et al. | [Retracted] Research on Application of Data Mining Algorithm in Cardiac Medical Diagnosis System |
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 |