CN109087702A - Four methods of diagnosis characterization information fusion method for the analysis of Chinese medicine health status - Google Patents
Four methods of diagnosis characterization information fusion method for the analysis of Chinese medicine health status Download PDFInfo
- Publication number
- CN109087702A CN109087702A CN201810878380.7A CN201810878380A CN109087702A CN 109087702 A CN109087702 A CN 109087702A CN 201810878380 A CN201810878380 A CN 201810878380A CN 109087702 A CN109087702 A CN 109087702A
- Authority
- CN
- China
- Prior art keywords
- information
- tester
- health status
- source
- analysis
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- 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
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- 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
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Public Health (AREA)
- Theoretical Computer Science (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Physics & Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Biology (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- Pathology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Medical Treatment And Welfare Office Work (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
For the four methods of diagnosis characterization information fusion method of Chinese medicine health status analysis, the information such as the prestige of clinical patient is acquired, hears, ask, cutting, the multi-source information for generating patient indicates, and marks its card type classification being subordinate to；The health status of tester is analyzed respectively using the characteristic present and its classification information of each information source, obtains multiple information sources to the decision-making assistant information of tester；Building information fusion model maximizes decision consistency, and the health status for returning to optimization analyzes result；The practical health status of contrast test person evaluates the performance of proposed algorithm with corresponding prediction result.It can detect the current health status of tester and lesion essence, enable itself bright physical condition of tester, provide reference to formulate intervention stratege.High-precision health status analysis can be provided as a result, providing foundation for health care.The four methods of diagnosis characterization information of clinical patient can be merged, more accurately and reliably state analysis result is obtained.
Description
Technical field
The present invention relates to Multi-label learnings, are especially for the four methods of diagnosis characterization information fusion of Chinese medicine health status analysis
Method.
Background technique
State is the logic starting point of Chinese medicine health community theory, health status refer to morphosis in the human body unit time,
Physiological function, psychological condition, the comprehensive state for adapting to external environment ability, embodiment is healthy situation and situation.Healthy shape
State analysis is by the prestige of acquisition, the information such as to hear using tcm theory as foundation, ask, cut and expressed with data mode, emphasize objectively
Human health status and lesion essence are evaluated, and the generality judgement (east Li Can Chinese medicine state [M] is provided to institute's illness, card
Beijing: China Traditional Chinese Medicine Publishing House, 2016).
Multi-label learning technology is used to handle the object in real world with ambiguity, in automatic image annotation, biology
The fields such as informatics, information retrieval and recommender system get the attention and apply.Specifically, the card of clinical patient
Often multimode is simultaneous holds under the arm for type distribution.So solving the problem analysis of Chinese medicine health status, multiple labeling based on artificial intelligence technology
Habit technology is introduced into the analysis of Chinese medicine health status.
According to the principle of Chinese medicine " ginseng is closed in the four methods of diagnosis ", state analysis is built upon on the basis of four methods of diagnosis information.In view of difference
Information source has otherness for the percentage contribution of prediction, and interrelated between different aforementioned sources, then passing through four methods of diagnosis method
The Global Information of clinical patient is collected, and then constructs information fusion model to analyze health status locating for the patient.
The features such as multimode state property and multiple labeling are presented in Chinese medicine health big data, so that traditional data analysis theories, side
Method and technology face the severe challenges such as validity, accuracy and computability.Therefore, research is for the analysis of Chinese medicine health status
Four methods of diagnosis characterization information fusion method is conducive to construct more accurate reliable identification model, is conducive to play artificial intelligence technology
Advantage promote cross discipline joint development and prosperity.
Summary of the invention
It is an object of the invention to hold under the arm for clinical patient multimode is simultaneous, and the source of diagnostic message has multiplicity
Property, the four methods of diagnosis characterization information fusion method for the analysis of Chinese medicine health status is provided.
The present invention the following steps are included:
1) it acquires the prestige of clinical patient, the information such as hears, asks, cutting, the multi-source information for generating patient indicates, and marks
Infuse its card type classification being subordinate to；
2) health status of tester is analyzed respectively using the characteristic present and its classification information of each information source,
Multiple information sources are obtained to the decision-making assistant information of tester；
3) building information fusion model maximizes decision consistency, for returning to the health status analysis knot of optimization
Fruit；
4) the practical health status of contrast test person evaluates the performance of proposed algorithm with corresponding prediction result.
In step 1), the prestige of the acquisition clinical patient such as hears, asks, cutting at the information, for generating the multi-source of patient
Information indicates, and the specific method for marking the card type classification that it is subordinate to can are as follows:
(1) four methods of diagnosis characterization information that clinical patient is extracted from electronic health record, forms information source A；Utilize observation instrument
Patient's tongue picture is obtained, tongue picture segmentation is realized based on U-Net network model, operator is then described using HSV, LAB and RGB and obtains tongue
As multiple character representations, information source B, information source C and information source D are separately constituted；
(2) health status of clinical patient is marked in doctor, is denoted as { l1,l2,...,lq, 1≤j≤q,
Middle ljFor j-th of card type of clinical patient, q is the sum of category label；
(3) algorithm is verified using ten folding cross validation methods: the ratio by the standardized data handled well according to 9 ︰ 1
Example is divided, and training data and test data are divided into.
In step 2), the characteristic present and its classification information using each information source is respectively to the health of tester
State is analyzed, and obtaining multiple information sources can to the specific method of the decision-making assistant information of tester are as follows:
(1) using the health status of SVM prediction tester, calculation formula are as follows:
Wherein,Indicate prediction result of i-th of tester about j-th of card type on data source A,It indicates i-th
Characteristic present information of the tester on data source A；
(2) union feature characterization and corresponding predictive information search the Top-k neighbour of tester, neighbour in training set
Select the similarity relationships based on tester and training sample, calculation formula are as follows:
Wherein,Similarity comprising tester and training sample on feature space, with cosine similarity method meter
It obtains；It is acquired by Jie Kade similarity method, the similarity comprising tester and training sample on label space；β
For threshold value, value range is [0,1]；
(3) similarity relationships sim is utilizedACorrelation modeling between verification type reconstructs the label space of tester:
Wherein,Indicate prediction result of i-th of tester about j-th of card type, Y on data source AzjIt indicates i-th
Actual value of z-th of the neighbour of tester in j-th of card type；
(4) step (1)~(3) are repeated, the state analysis result based on information source B~D is respectively obtained.
In step 3), the building information fusion model maximizes decision consistency, for returning to the strong of optimization
The specific method of health state analysis result can are as follows:
(1) the final knot of tester is obtained using multiple state outcomes of clinical patient four methods of diagnosis characterization information prediction
Fruit constructs following optimization object function and is solved:
Wherein,Indicate that optimum results of i-th of tester in j-th of card type, the optimum results pass through fusion
The decision information of multi-source obtains, W={ w1,w2,...,wMBe M information source weight distribution, wherein M=4,
In addition, cmIndicate the set of (i, j), and (i, j) meetsα is threshold value, and value range is [0,1]；
(2) weight is initialized, is enabledSetting:
(3) fixed W solves Y using gradient descent method*, calculation formula are as follows:
(4) fixed Y*, W, calculation formula are solved using method of Lagrange multipliers are as follows:
(5) step (3) and (4) are repeated, until optimization aim convergence, returns to the optimum results Y of tester's health status*。
In step 4), the practical health status of the contrast test person evaluates mentioned algorithm with corresponding prediction result
The specific method of performance can are as follows:
It is predicted using category label of the mentioned method to tester in test data, and uses following five indexs pair
The performance of mentioned algorithm is evaluated:
(1) Hamming loses: for investigating misclassification situation of the sample on single marking, the evaluation index is the smaller the better；
(2) 1- error rate: for investigating in the category label collating sequence of sample, the label of sequence front end is not belonging to
The case where mark of correlation set, the evaluation index are the smaller the better；
(3) it coverage rate: for investigating in the category label collating sequence of sample, covers and is searched needed for all mark of correlation
Rope depth profile, the evaluation index are the smaller the better；
(4) sequence loss: there is the case where misordering in the category label collating sequence of sample for investigating, this is commented
Valence index is the smaller the better；
(5) mean accuracy: for investigating in the category label collating sequence of sample, the label before mark of correlation is come
The case where being still mark of correlation, which is the bigger the better.
Compared with prior art, the present invention is capable of detecting when the current health status of tester and lesion essence, so that surveying
The physical condition of itself can be illustrated in examination person, provide reference to formulate intervention stratege.
The present invention is capable of providing high-precision health status analysis as a result, providing foundation for health care.
The present invention can merge the four methods of diagnosis characterization information of clinical patient, to obtain more accurately and reliably state point
Analyse result.
Detailed description of the invention
Fig. 1 is the schematic diagram of tongue picture segmentation.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
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.
The present embodiment includes the following steps:
1) it acquires the prestige of 729 clinical patients, the information such as hears, asks, cutting, for generating the multi-source information table of patient
Show, and mark its card type classification being subordinate to, amounts to 339 card type classification numbers；
(1) four methods of diagnosis characterization information that clinical patient is extracted from electronic health record, forms information source A；Utilize observation instrument
Patient's tongue picture is obtained, tongue picture segmentation is realized based on U-Net network model, as shown in Figure 1.Then it is described using HSV, LAB and RGB
Operator obtains the multiple character representations of tongue picture, separately constitutes information source B, information source C and information source D；
(2) health status of clinical patient is marked in doctor, is denoted as { l1,l2,...,lq}(1≤j≤q).Its
Middle ljFor j-th of card type of clinical patient, q is the sum of category label；
(3) algorithm is verified using ten folding cross validation methods: the ratio by the standardized data handled well according to 9 ︰ 1
Example is divided, and training data and test data are divided into.
2) health status for being analyzed tester respectively using the characteristic present and its classification information of each information source, is obtained more
Decision-making assistant information of a information source to tester；
(1) tester's health status, calculation formula are predicted using SVM are as follows:
Wherein,Indicate prediction result of i-th of tester about j-th of card type on data source A,It indicates i-th
Characteristic present information of the tester on data source A；
(2) union feature characterization and corresponding predictive information search the Top-k neighbour of tester in training set.Neighbour
Select the similarity relationships based on tester and training sample, calculation formula are as follows:
Wherein,Similarity comprising tester and training sample on feature space, with cosine similarity method meter
It obtains；It is acquired by Jie Kade similarity method, the similarity comprising tester and training sample on label space；β
For threshold value, value range is [0,1]；
(3) similarity relationships sim is utilizedACorrelation modeling between verification type reconstructs the label space of tester:
Wherein,Indicate prediction result of i-th of tester about j-th of card type, Y on data source AzjIt indicates i-th
Actual value of z-th of the neighbour of tester in j-th of card type；
(4) by the prediction result on information source A respectively with BSVM (M.R.Boutell, J.Luo, X.Shen,
C.M.Brown,Learning multi-label scene classification,Pattern Recognition,2004,
37 (9): 1757-1771) and LIFT (M.Zhang, L.Wu, LIFT:Multi-label learning with label-
specific features,IEEE Transactions on Pattern Analysis and Machine
Intelligence, 2015,37 (1): 107-120) method is compared, and experimental result is as shown in table 1.Algorithm 1 is corresponding to be
The verification result of the mentioned algorithm of the present invention；It is the verification result of LIFT that algorithm 2 is corresponding；It is the verifying of BSVM that algorithm 3 is corresponding
As a result.From table 1 it follows that the present invention is by considering that label correlated performance is better than other calculations in most evaluation index
Method.
Table 1
| Algorithm | Hamming loss | 1- error rate | Coverage rate | Sequence loss | Mean accuracy |
| 1 | 0.0122±0.0008 | 0.1701±0.0330 | 0.3251±0.0435 | 0.0617±0.0081 | 0.6405±0.0265 |
| 2 | 0.0133±0.0009 | 0.1990±0.0377 | 0.4168±0.0408 | 0.0882±0.0085 | 0.5910±0.0260 |
| 3 | 0.0123±0.0009 | 0.1358±0.0263 | 0.3453±0.0436 | 0.0665±0.0076 | 0.6355±0.0177 |
(5) step (1)~(3) are repeated, the health status analysis result based on information source B~D is respectively obtained.
3) building information fusion model maximizes decision consistency, for returning to the health status analysis knot of optimization
Fruit；
(1) the multiple state identification results predicted using clinical patient four methods of diagnosis characterization information are final to obtain tester
Solved as a result, constructing following optimization object function:
Wherein,Indicate optimum results of i-th of tester in j-th of card type, which passes through fusion multi-source
Decision information obtain.W={ w1,w2,...,wMBe M information source weight distribution (M=4 here),Separately
Outside, cmIndicate the set of (i, j), and (i, j) meetsα is threshold value, and value range is [0,1]；
(2) weight is initialized.It enablesSetting:
(3) fixed W solves Y using gradient descent method*, calculation formula are as follows:
(4) fixed Y*, W, calculation formula are solved using method of Lagrange multipliers are as follows:
(5) step (3)~(4) are repeated, until optimization aim convergence, return to the optimum results Y of tester's health status*。
4) it is analyzed using health status of the mentioned method to tester in test data；
Mentioned algorithm is compared with the prediction result of each information source, as shown in table 2.From table 2 it can be seen that being mentioned
Algorithm can obtain optimal result by fuse information source A~D in most of evaluation index.
Table 2
| Hamming loss | 1- error rate | Coverage rate | Sequence loss | Mean accuracy | |
| Information source A | 0.0122±0.0010 | 0.1701±0.0330 | 0.3251±0.0435 | 0.0617±0.0081 | 0.6405±0.0265 |
| Information source B | 0.0180±0.0014 | 0.5624±0.0657 | 0.4090±0.0443 | 0.0978±0.0071 | 0.3536±0.0203 |
| Information source C | 0.0181±0.0015 | 0.6090±0.0470 | 0.4122±0.0412 | 0.0982±0.0058 | 0.3447±0.0160 |
| Information source D | 0.0181±0.0015 | 0.5968±0.0595 | 0.4075±0.0413 | 0.0968±0.0083 | 0.3516±0.0241 |
| Information fusion | 0.0118±0.0011 | 0.1604±0.0272 | 0.3328±0.0634 | 0.0637±0.0108 | 0.6473±0.0191 |
Mentioned algorithm is compared with other blending algorithms, as shown in table 3.Corresponding algorithm 1 is that the present invention proposes calculation
The verification result of method；The corresponding average result based on the prediction of all information sources of algorithm 2；Algorithm 3 is corresponding to be based on all information
All information sources are connected, are then classified using SVM by the voting results of source prediction, algorithm 4.It can from table 3
Out, the mentioned algorithm of the present invention has optimal result.
Table 3
| Algorithm | Hamming loss | 1- error rate | Coverage rate | Sequence loss | Mean accuracy |
| 1 | 0.0118±0.0011 | 0.1604±0.0272 | 0.3328±0.0634 | 0.0637±0.0108 | 0.6473±0.0191 |
| 2 | 0.0161±0.0019 | 0.2386±0.0350 | 0.3716±0.0578 | 0.0778±0.0102 | 0.5222±0.0190 |
| 3 | 0.0177±0.0020 | 0.4073±0.0621 | 0.3715±0.0578 | 0.0803±0.0101 | 0.4422±0.0242 |
| 4 | 0.0123±0.0007 | 0.1427±0.0409 | 0.3473±0.0427 | 0.0673±0.0083 | 0.6343±0.0158 |
The present invention first pre-processes the information of four methods of diagnosis Acquisition Instrument capture, then analyzes the pre- of each information source respectively
Result is surveyed to judge that the health status of tester, the prediction result for finally merging multiple characteristic present information make state identification
Consistency maximizes, and accurately and reliably refers to formulate intervention stratege for tester and provide.
Claims (5)
1. the four methods of diagnosis characterization information fusion method for the analysis of Chinese medicine health status, it is characterised in that the following steps are included:
1) prestige of clinical patient is acquired, hears, ask, cutting information, the multi-source information for generating patient indicates, and marks its person in servitude
The card type classification of category；
2) health status of tester is analyzed respectively using the characteristic present and its classification information of each information source, is obtained
Decision-making assistant information of multiple information sources to tester；
3) building information fusion model maximizes decision consistency, and the health status for returning to optimization analyzes result；
4) the practical health status of contrast test person evaluates the performance of proposed algorithm with corresponding prediction result.
2. the four methods of diagnosis characterization information fusion method for the analysis of Chinese medicine health status as described in claim 1, it is characterised in that
In step 1), the prestige of the acquisition clinical patient is heard, asks, cutting information, and the multi-source information for generating patient indicates, and
Mark the card type classification that it is subordinate to method particularly includes:
(1) four methods of diagnosis characterization information that clinical patient is extracted from electronic health record, forms information source A；It is obtained using observation instrument
Patient's tongue picture realizes tongue picture segmentation based on U-Net network model, and it is more then to describe operator acquisition tongue picture using HSV, LAB and RGB
A character representation separately constitutes information source B, information source C and information source D；
(2) health status of clinical patient is marked in doctor, is denoted as { l1,l2,...,lq, 1≤j≤q, wherein ljFor
J-th of card type of clinical patient, q are the sum of category label；
(3) algorithm is verified using ten folding cross validation methods: by the standardized data handled well according to 9 ︰ 1 ratio into
Row divides, and is divided into training data and test data.
3. the four methods of diagnosis characterization information fusion method for the analysis of Chinese medicine health status as described in claim 1, it is characterised in that
In step 2), the characteristic present and its classification information using each information source respectively divides the health status of tester
Analysis obtains multiple information sources to the decision-making assistant information of tester method particularly includes:
(1) using the health status of SVM prediction tester, calculation formula are as follows:
Wherein,Indicate prediction result of i-th of tester about j-th of card type on data source A,Indicate i-th of test
Characteristic present information of the person on data source A；
(2) union feature characterization and corresponding predictive information search the Top-k neighbour of tester, neighbour's selection in training set
Similarity relationships based on tester and training sample, calculation formula are as follows:
Wherein,Similarity comprising tester and training sample on feature space is calculated with cosine similarity method
It arrives；It is acquired by Jie Kade similarity method, the similarity comprising tester and training sample on label space；β is threshold
Value, value range are [0,1]；
(3) similarity relationships sim is utilizedACorrelation modeling between verification type reconstructs the label space of tester:
Wherein,Indicate prediction result of i-th of tester about j-th of card type, Y on data source AzjIndicate i-th of test
Actual value of z-th of the neighbour of person in j-th of card type；
(4) step (1)~(3) are repeated, the state analysis result based on information source B~D is respectively obtained.
4. the four methods of diagnosis characterization information fusion method for the analysis of Chinese medicine health status as described in claim 1, it is characterised in that
In step 3), the building information fusion model maximizes decision consistency, for returning to the health status analysis of optimization
As a result method particularly includes:
(1) obtained using multiple state outcomes of the clinical patient four methods of diagnosis characterization information prediction tester it is final as a result,
Following optimization object function is constructed to be solved:
Wherein,Indicate that optimum results of i-th of tester in j-th of card type, the optimum results pass through fusion multi-source
Decision information obtain, W={ w1,w2,...,wMBe M information source weight distribution, wherein M=4,Separately
Outside, cmIndicate the set of (i, j), and (i, j) meetsα is threshold value, and value range is [0,1]；
(2) weight is initialized, is enabledSetting:
(3) fixed W solves Y using gradient descent method*, calculation formula are as follows:
(4) fixed Y*, W, calculation formula are solved using method of Lagrange multipliers are as follows:
(5) step (3) and (4) are repeated, until optimization aim convergence, returns to the optimum results Y of tester's health status*。
5. the four methods of diagnosis characterization information fusion method for the analysis of Chinese medicine health status as described in claim 1, it is characterised in that
In step 4), the practical health status of the contrast test person evaluates the tool of the performance of mentioned algorithm with corresponding prediction result
Body method are as follows:
It is predicted using category label of the mentioned method to tester in test data, and using following five indexs to being mentioned
The performance of algorithm is evaluated:
(1) Hamming loses: for investigating misclassification situation of the sample on single marking, the evaluation index is the smaller the better；
(2) 1- error rate: for investigating in the category label collating sequence of sample, the label of sequence front end is not belonging to correlation
The case where tag set, the evaluation index are the smaller the better；
(3) coverage rate: for investigating in the category label collating sequence of sample, search needed for covering all mark of correlation is deep
Situation is spent, the evaluation index is the smaller the better；
(4) sequence loss: there is the case where misordering in the category label collating sequence of sample for investigating, which refers to
It marks the smaller the better；
(5) mean accuracy: for investigating in the category label collating sequence of sample, the label before coming mark of correlation is still
The case where mark of correlation, the evaluation index are the bigger the better.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810878380.7A CN109087702B (en) | 2018-08-03 | 2018-08-03 | Four-diagnosis representation information fusion method for traditional Chinese medicine health state analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810878380.7A CN109087702B (en) | 2018-08-03 | 2018-08-03 | Four-diagnosis representation information fusion method for traditional Chinese medicine health state analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109087702A true CN109087702A (en) | 2018-12-25 |
| CN109087702B CN109087702B (en) | 2021-07-16 |
Family
ID=64833581
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810878380.7A Active CN109087702B (en) | 2018-08-03 | 2018-08-03 | Four-diagnosis representation information fusion method for traditional Chinese medicine health state analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109087702B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111260619A (en) * | 2020-01-14 | 2020-06-09 | 浙江中医药大学 | Tongue body automatic segmentation method based on U-net model |
| CN112200091A (en) * | 2020-10-13 | 2021-01-08 | 深圳市悦动天下科技有限公司 | Tongue region identification method and device and computer storage medium |
| CN112530584A (en) * | 2020-12-15 | 2021-03-19 | 贵州小宝健康科技有限公司 | Medical diagnosis assisting method and system |
| CN113409938A (en) * | 2021-06-30 | 2021-09-17 | 海南医学院 | Modeling method and system of traditional Chinese medicine syndrome type prediction model of systemic lupus erythematosus |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101129261A (en) * | 2007-02-09 | 2008-02-27 | 北京中医药大学 | Device and method for acquiring and recognizing pulsation information and tongue inspection information |
| CN101647696A (en) * | 2008-08-13 | 2010-02-17 | 上海经路通中医药科技发展有限公司 | Intelligent system for health diagnosis and treatment |
| US20100226856A1 (en) * | 2009-03-06 | 2010-09-09 | Franco Vitaliano | Dynamic bio-nanoparticle elements |
| CN104766068A (en) * | 2015-04-20 | 2015-07-08 | 江西中医药大学 | Random walk tongue image extraction method based on multi-rule fusion |
| CN105528529A (en) * | 2016-02-20 | 2016-04-27 | 成都中医药大学 | Data processing method of traditional Chinese medicine clinical skill evaluation system based on big data analysis |
| CN106874655A (en) * | 2017-01-16 | 2017-06-20 | 西北工业大学 | Traditional Chinese medical science disease type classification Forecasting Methodology based on Multi-label learning and Bayesian network |
| CN108198621A (en) * | 2018-01-18 | 2018-06-22 | 中山大学 | A kind of database data synthesis dicision of diagnosis and treatment method based on neural network |
-
2018
- 2018-08-03 CN CN201810878380.7A patent/CN109087702B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101129261A (en) * | 2007-02-09 | 2008-02-27 | 北京中医药大学 | Device and method for acquiring and recognizing pulsation information and tongue inspection information |
| CN101647696A (en) * | 2008-08-13 | 2010-02-17 | 上海经路通中医药科技发展有限公司 | Intelligent system for health diagnosis and treatment |
| US20100226856A1 (en) * | 2009-03-06 | 2010-09-09 | Franco Vitaliano | Dynamic bio-nanoparticle elements |
| CN104766068A (en) * | 2015-04-20 | 2015-07-08 | 江西中医药大学 | Random walk tongue image extraction method based on multi-rule fusion |
| CN105528529A (en) * | 2016-02-20 | 2016-04-27 | 成都中医药大学 | Data processing method of traditional Chinese medicine clinical skill evaluation system based on big data analysis |
| CN106874655A (en) * | 2017-01-16 | 2017-06-20 | 西北工业大学 | Traditional Chinese medical science disease type classification Forecasting Methodology based on Multi-label learning and Bayesian network |
| CN108198621A (en) * | 2018-01-18 | 2018-06-22 | 中山大学 | A kind of database data synthesis dicision of diagnosis and treatment method based on neural network |
Non-Patent Citations (2)
| Title |
|---|
| JIA ZHANG.ETC: "Computational drug repositioning using collaborative filtering via multi-source fusion", 《EXPERT SYSTEMS WITH APPLICATIONS》 * |
| 杨柳，等: "面向认知的多源数据学习理论和算法研究进展", 《软件学报》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111260619A (en) * | 2020-01-14 | 2020-06-09 | 浙江中医药大学 | Tongue body automatic segmentation method based on U-net model |
| CN112200091A (en) * | 2020-10-13 | 2021-01-08 | 深圳市悦动天下科技有限公司 | Tongue region identification method and device and computer storage medium |
| CN112530584A (en) * | 2020-12-15 | 2021-03-19 | 贵州小宝健康科技有限公司 | Medical diagnosis assisting method and system |
| CN113409938A (en) * | 2021-06-30 | 2021-09-17 | 海南医学院 | Modeling method and system of traditional Chinese medicine syndrome type prediction model of systemic lupus erythematosus |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109087702B (en) | 2021-07-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110993064B (en) | Deep learning-oriented medical image labeling method and device | |
| KR101977174B1 (en) | Apparatus, method and computer program for analyzing image | |
| CN109087702A (en) | Four methods of diagnosis characterization information fusion method for the analysis of Chinese medicine health status | |
| CN105009171B (en) | For learning the method and apparatus split automatically based on collection of illustrative plates of enhancing | |
| Sainju et al. | Automated bleeding detection in capsule endoscopy videos using statistical features and region growing | |
| Bjornsson et al. | Associative image analysis: a method for automated quantification of 3D multi-parameter images of brain tissue | |
| CN110491502B (en) | Microscope video stream processing method, system, computer device and storage medium | |
| CN109770932A (en) | The processing method of multi-modal brain neuroblastoma image feature | |
| CN105205501B (en) | A kind of weak mark image object detection method of multi classifier combination | |
| CN110018524A (en) | A kind of X-ray safety check contraband recognition methods of view-based access control model-attribute | |
| CN106951825A (en) | A kind of quality of human face image assessment system and implementation method | |
| CN103745217A (en) | Automatic analysis method of tongue color and coating color in traditional Chinese medicine based on image retrieval | |
| KR102045223B1 (en) | Apparatus, method and computer program for analyzing bone age | |
| CN109770903A (en) | The classification prediction technique of functional magnetic resonance imaging, system, device | |
| CN107016395A (en) | Primary brain lymthoma and the discrimination method of glioblastoma based on rarefaction representation system | |
| CN102436551B (en) | Computer aided gastric cancer diagnosis method based on target tracking | |
| CN104112292A (en) | Positioning Anatomical Landmarks In Volume Data Sets | |
| CN112419452B (en) | Rapid merging system and method for PD-L1 digital pathological section images of stomach cancer | |
| CN106778047A (en) | A kind of traditional Chinese medical science facial diagnosis integrated system based on various dimensions medical image | |
| CN114600155A (en) | Weakly supervised multitask learning for cell detection and segmentation | |
| CN115546605A (en) | Training method and device based on image labeling and segmentation model | |
| CN103839048B (en) | Stomach CT image lymph gland recognition system and method based on low-rank decomposition | |
| CN110503147A (en) | Multi-mode image categorizing system based on correlation study | |
| Weng et al. | Multi-label symptom analysis and modeling of TCM diagnosis of hypertension | |
| Ahmad et al. | Brain tumor detection & features extraction from MR images using segmentation, image optimization & classification techniques |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |