CN109887603A - A kind of computer-assisted medical data processing system and method - Google Patents
A kind of computer-assisted medical data processing system and method Download PDFInfo
- Publication number
- CN109887603A CN109887603A CN201910043458.8A CN201910043458A CN109887603A CN 109887603 A CN109887603 A CN 109887603A CN 201910043458 A CN201910043458 A CN 201910043458A CN 109887603 A CN109887603 A CN 109887603A
- Authority
- CN
- China
- Prior art keywords
- patient
- information
- frequent
- data
- carries out
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
Landscapes
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention belongs to medical ancillary technique field, a kind of computer-assisted medical data processing system and method are disclosed；Input system is carried out to the data information of patient, the information of patient is transmitted in Cloud Server store by wireless signal transceiver by computer to be shared；Classified storage is carried out in reservoir to the data information of patient, timely the information of patient is updated；When data information of the clinician to patient is inquired, patient information is verified by the name or certificate that input patient, the information of patient is inquired, carries out relevant diagnoses and treatment；The related data information of patient is shown by display, is checked for clinician.The present invention effectively improves SVM algorithm to the classification performance of minority class, improves operational efficiency；Digging efficiency and scalability are improved, is allowed to be suitable for the big data environment that dynamic increases；It solves the contradiction that matching primitives speed and precision cannot get both, realizes the correct matching of characteristic point well.
Description
Technical field
The invention belongs to medical ancillary technique field more particularly to a kind of computer-assisted medical data processing system and sides
Method.
Background technique
Computer-aided diagnosis or computer aided detection refer to by iconography, Medical Image Processing and other
Possible physiology, biochemical apparatus, in conjunction with the analytical calculation of computer, auxiliary discovery lesion improves the accuracy rate of diagnosis.Often now
The cad technique said is primarily referred to as the computer aided technique based on Medical Imaging.With the computer aided detection (CAD) phase
Difference, the latter's emphasis are detections, and computer is only needed to be labeled abnormal sign, be carried out at common image on this basis
Reason, and without further being diagnosed.That is, computer-aided diagnosis is the extension and final purpose of computer aided detection, phase
Ying Di, computer aided detection are basis and the only stage which must be passed by of computer-aided diagnosis.Cad technique is otherwise known as " the of doctor
Three eyes ", the extensive use of CAD system help to improve the sensibility and specificity of diagnosis.
In conclusion problem of the existing technology is:
(1) existing computer-assisted medical data processing system, inquires that specific patient information efficiency is lower and patient believes
Breath is unable to real-time update.
It (2), cannot during carrying out classified storage using relevant information of traditional algorithm to patient in the prior art
SVM algorithm is effectively improved to the classification performance of minority class, reduces operational efficiency.
(3) process to be timely updated in the prior art using traditional algorithm to the data information of the patient in whole system
In, digging efficiency and scalability cannot be effectively improved, is allowed to not be suitable for the big data environment that dynamic increases.
(4) it is existing during clinician verifies patient information by the name or certificate for inputting patient
Information Authentication algorithm not can solve the contradiction that matching primitives speed and precision cannot get both, can not realize characteristic point well
Correct matching.
Summary of the invention
In view of the problems of the existing technology, the present invention provides a kind of computer-assisted medical data processing system and sides
Method.
The invention is realized in this way a kind of computer aided medicine data processing method, the computer aided medicine
Data processing method the following steps are included:
The first step carries out input system to the data information of patient, and computer is by wireless signal transceiver by patient's
Information, which is transmitted in Cloud Server store, to be shared；
Second step carries out classified storage in reservoir to the data information of patient, timely carries out to the information of patient
It updates；
Third step, when data information of the clinician to patient is inquired, by input patient name or
Certificate verifies patient information, inquires the information of patient, carries out relevant diagnoses and treatment；
The related data information of 4th step, patient is shown by display, is checked for clinician.
Further, classified storage is carried out to the relevant information of patient, using improved unbalanced data SVM algorithm, packet
Include following steps:
Step 1, is arranged the most class sample point number MajorN=m*Minor N sampled in advance, and m is that the two is quantitative
Ratio；
Step 2 establishes the similar matrix S ∈ R based on Gaussian kernel first with most class sample setsn×n, n is former
There is the number of most class samples；
Step 3 carries out clustering to most class samples using above-mentioned spectral clustering, generates cluster A1..., Ak, k
=MajorN；
Step 4 selects sample point representative in each cluster, wherein the sample choosing in each cluster
The size that number depends on the average distance of the size of the cluster and the sample in the cluster and minority class sample point is selected, cluster is got over
Greatly, the sample number of selection is more；More close from minority class sample, what is selected is fewer；
KDistil=K (xi, xi)+K(xl, xl)-2K(xi, xl)；
Wherein KsizeiTo cluster AiSize, IDistiTo cluster AiTo the average distance of minority class,It is every
The number of samples selected in one cluster；
Step 5, for each cluster, before selectionIt is a from the smallest sample of minority class sample point average distance
New most class training sample subsets are formed after this combination；
Step 6, most class training sample subsets that sampling is obtained and whole minority class sample group cooperations are new instruction
Practice sample, is input in SVM algorithm and is trained study, wherein nuclear parameter is identical as the nuclear parameter of similar matrix S, i.e. spectral clustering
It is all carried out in a space with svm classifier algorithm；
Step 7 is recognized according to the classification that the classification interface that training obtains carries out new samples.
Further, it timely updates to the data information of patient, using improved data more new algorithm, detailed process is as follows:
Original transaction database is set as DB, transaction set T={ T1, T2..., Tn, candidate CD, frequent item set is
LD, original transaction Database size is | DB |；Newly-increased database is db, frequent item set Ld, newly-increased Database size is | db |；
Updated database is S, S=DB ∪ db, frequent item set LS, support sup；
Input: original transaction data set DB；Newly-increased transaction data set (TDS) db；
Output: updated data set S；Frequent item set LS；
Step 1 scans raw data set, obtains 1 item collection of candidate of DBFrequent 1 item collection is obtained according to supportEstablished after sequence, FList is grouped according to grouping strategy, construct the frequent pattern tree (fp tree) of DB, to frequent pattern tree (fp tree) into
The excavation of row frequent item set obtains the frequent item set L of DBD；
Step 2 scans newly-increased data set, obtains 1 item collection of candidate of dbIn read step 1Merge
1 item collection of candidate of data set S after to updateFrequent 1 item collection of S is obtained according to supportIt is established after sequence
FList ' is grouped according to grouping strategy, constructs the frequent pattern tree (fp tree) of db；
Step 3, to frequent pattern tree (fp tree) carry out frequent item set excavation, while in read step 1 DB frequent item set LD,
To each frequent mode k and L of generationDIt compares: if k belongs to LD, then k is the frequent episode of original data set, by the support of k with
In LDCorresponding support counting addition can obtain support counting of the k in S, if tale is greater than the minimum support meter of S
Number, then be added the part frequent item set L ' of S, and from LDIt is middle to delete this；If k is not belonging to LD, then k is the frequent of newly-increased data set
, it whether frequently not to can determine that in S, is added into the Candidate Set C of SS；Differentiate LDIn remaining item, if its support counting
Minimum support greater than S counts, then L ' is added；
Step 4 scans raw data set, the Candidate Set C of S in read step threeS, judge whether it is the frequent item set of S,
Frequent item set is merged with part frequent item set L ', the frequent item set L of data set S after just being updatedS。
Another object of the present invention is to provide a kind of calculating for executing the computer aided medicine data processing method
Machine medical assistance data processing system, the computer-assisted medical data processing system include:
MIM message input module is connect with central processing module, carries out typing information to the related data information of patient；
Clinician's enquiry module, connect with central processing module, and according to the identity information of patient, clinician carries out disease
The reading of people's information carries out subsequent treatment；
Display module is connect with central processing module, by the related data information for showing patient using display；
Transmission module is exported, is connect with central processing module, by being believed the data of patient using wireless signal transmitter
Breath is transmitted in Cloud Server, carries out access reference；
Cloud Server is connect with central processing module, provides relevant case for clinician, and to the information of patient
It is stored, is consulted for reference for clinicians；
Storage module is connect with central processing module, carries out classified storage to the relevant information of patient；
Information updating module is connect with central processing module, is timely updated to the data information of the patient in system；
Patient information authentication module, connect with central processing module, the name or certificate that clinician passes through input patient
Patient information is verified.
It is auxiliary using the medical treatment of the calculation machine medical assistance data processing method that another object of the present invention is to provide a kind of
Help platform.
Advantages of the present invention and good effect are as follows: storage module carries out classified storage to the relevant information of patient in the present invention
During, in order to effectively improve SVM algorithm to the classification performance of minority class, operational efficiency is improved, using a kind of improved
Unbalanced data SVM algorithm.
During information updating module timely updates to the data information of the patient in whole system in the present invention, in order to
Digging efficiency and scalability are improved, is allowed to be suitable for the big data environment that dynamic increases, be updated using a kind of improved data
Algorithm.
The name or certificate that clinician passes through input patient in patient information authentication module in the present invention are to patient information
During being verified, in order to solve the contradiction that matching primitives speed and precision cannot get both, in order to realize well
The correct matching of characteristic point, with high use value, using a kind of improved data characteristics matching algorithm.
Detailed description of the invention
Fig. 1 is computer aided medicine data processing method flow chart provided in an embodiment of the present invention.
Fig. 2 is the structural schematic diagram of computer-assisted medical data processing system provided in an embodiment of the present invention；
In figure: 1, MIM message input module；2, clinician's enquiry module；3, display module；4, central processing module；5, defeated
Transmission module out；6, Cloud Server；7, storage module；8, information updating module；9, patient information authentication module.
Specific embodiment
In order to further understand the content, features and effects of the present invention, the following examples are hereby given, and cooperate attached drawing
Detailed description are as follows.
Structure of the invention is explained in detail with reference to the accompanying drawing.
As shown in Figure 1, computer aided medicine data processing method provided in an embodiment of the present invention the following steps are included:
S101: firstly, carrying out input system to the data information of patient, computer passes through wireless signal transceiver for patient
Information be transmitted in Cloud Server that carry out storage shared；
S102: classified storage is carried out in reservoir to the data information of patient, timely the information of patient is carried out more
Newly；
S103: when data information of the clinician to patient is inquired, by the name or card that input patient
Part verifies patient information, inquires the information of patient, carries out relevant diagnoses and treatment；
S104: the related data information of patient is shown by display, is checked for clinician.
As shown in Fig. 2, computer-assisted medical data processing system provided in an embodiment of the present invention includes: information input mould
Block 1, clinician's enquiry module 2, display module 3, central processing module 4, output transmission module 5, Cloud Server 6, storage mould
Block 7, information updating module 8, patient information authentication module 9.
MIM message input module 1 is connect with central processing module 4, carries out typing information to the related data information of patient；
Clinician's enquiry module 2 is connect with central processing module 4, and according to the identity information of patient, clinician is carried out
The reading of patient information carries out subsequent treatment；
Display module 3 is connect with central processing module 4, by the related data information for showing patient using display；
Transmission module 5 is exported, is connect with central processing module 4, by utilizing wireless signal transmitter by the data of patient
Information is transmitted in Cloud Server, carries out access reference；
Cloud Server 6 is connect with central processing module 4, provides relevant case for clinician, and to the letter of patient
Breath is stored, and is consulted for reference for clinicians；
Storage module 7 is connect with central processing module 4, carries out classified storage to the relevant information of patient；
Information updating module 8 is connect with central processing module 4, in time more to the data information of the patient in whole system
Newly；
Patient information authentication module 9 is connect with central processing module 4, the name or card that clinician passes through input patient
Part verifies patient information.
During the storage module 7 carries out classified storage to the relevant information of patient, calculated in order to effectively improve SVM
Method improves operational efficiency to the classification performance of minority class, using a kind of improved unbalanced data SVM algorithm, including it is following
Step:
Step 1, is arranged the most class sample point number MajorN=m*Minor N sampled in advance, and m is that the two is quantitative
Ratio；
Step 2 establishes the similar matrix S ∈ R based on Gaussian kernel first with most class sample setsn×n, n is former
There is the number of most class samples；
Step 3 carries out clustering to most class samples using above-mentioned spectral clustering, generates cluster A1..., Ak, k
=MajorN；
Step 4 selects sample point representative in each cluster, wherein the sample choosing in each cluster
The size that number depends on the average distance of the size of the cluster and the sample in the cluster and minority class sample point is selected, cluster is got over
Greatly, the sample number of selection is more；More close from minority class sample, what is selected is fewer；
So selection be in order to purposefully delete the boundary sample information point in most classes, it is specific as follows:
KDistil=K (xi, xi)+K(xl, xl)-2K(xi, xl)；
Wherein KsizeiTo cluster AiSize, IDistiTo cluster AiTo the average distance of minority class,It is every
The number of samples selected in one cluster；
Step 5, for each cluster, before selectionIt is a from the smallest sample of minority class sample point average distance
New most class training sample subsets are formed after this combination；
Step 6, most class training sample subsets that sampling is obtained and whole minority class sample group cooperations are new instruction
Practice sample, is input in SVM algorithm and is trained study, wherein nuclear parameter is identical as the nuclear parameter of similar matrix S, i.e. spectral clustering
It is all carried out in a space with svm classifier algorithm；
Step 7 is recognized according to the classification that the classification interface that training obtains carries out new samples.
During the information updating module 8 timely updates to the data information of the patient in whole system, in order to mention
High digging efficiency and scalability are allowed to be suitable for the big data environment that dynamic increases, are updated and calculated using a kind of improved data
Method, detailed process is as follows:
Original transaction database is set as DB, transaction set T={ T1, T2..., Tn, candidate CD, frequent item set is
LD, original transaction Database size is | DB |；Newly-increased database is db, frequent item set Ld, newly-increased Database size is | db |；
Updated database is S, S=DB ∪ db, frequent item set LS, support sup；
Input: original transaction data set DB；Newly-increased transaction data set (TDS) db；
Output: updated data set S；Frequent item set LS；
Step 1 scans raw data set, obtains 1 item collection of candidate of DBFrequent 1 item collection is obtained according to supportEstablished after sequence, FList is grouped according to grouping strategy, construct the frequent pattern tree (fp tree) of DB, to frequent pattern tree (fp tree) into
The excavation of row frequent item set obtains the frequent item set L of DBD；
Step 2 scans newly-increased data set, obtains 1 item collection of candidate of dbIn read step 1Merge
1 item collection of candidate of data set S after to updateFrequent 1 item collection of S is obtained according to supportIt is established after sequence
FList ' is grouped according to grouping strategy, constructs the frequent pattern tree (fp tree) of db；
Step 3, to frequent pattern tree (fp tree) carry out frequent item set excavation, while in read step 1 DB frequent item set LD,
To each frequent mode k and L of generationDIt compares: if k belongs to LD, then k is the frequent episode of original data set, by the support of k with
In LDCorresponding support counting addition can obtain support counting of the k in S, if tale is greater than the minimum support meter of S
Number, then be added the part frequent item set L ' of S, and from LDIt is middle to delete this；If k is not belonging to LD, then k is the frequent of newly-increased data set
, it whether frequently not to can determine that in S, is added into the Candidate Set C of SS；Differentiate LDIn remaining item, if its support counting
Minimum support greater than S counts, then L ' is added；
Step 4 scans raw data set, the Candidate Set C of S in read step threeS, judge whether it is the frequent item set of S,
Frequent item set is merged with part frequent item set L ', the frequent item set L of data set S after just being updatedS。
Clinician carries out patient information by the name or certificate of input patient in the patient information authentication module 9
During verifying, in order to solve the contradiction that matching primitives speed and precision cannot get both, in order to realize feature well
The correct matching of point, using a kind of improved data characteristics matching algorithm, specifically includes following with high use value
Step:
Step 1 selectes the rectangular area comprising target as object matching template by hand in initial frame, passes through M-
The Feature Descriptor of SURF algorithm calculation template；
Step 2 carries out divided-fit surface in new frame, calculates the M-SURF feature key points of every piece of template；
Step 3 obtains matching double points in Euclidean space using the method for arest neighbors and time neighbour's ratio；
Step 4 rejects pseudo- match point using LMedS method；
Step 5, using SFM algorithm to plane characteristic point to progress Stereo matching.
The working principle of the invention is:
Clinician's enquiry module 2 is according to the identity information of patient, and clinician carries out the reading of patient information, after progress
Continuous treatment；Export transmission module 5 by using wireless signal transmitter by the data information transfer of patient into Cloud Server,
Carry out access reference；Cloud Server 6 provides relevant case for clinician, and stores to the information of patient, for facing
Bed doctor is with reference to access；Storage module 7 carries out classified storage to the relevant information of patient；Information updating module 8 is to whole system
In the data information of patient timely update；The name or card that clinician passes through input patient in patient information authentication module 9
Part verifies patient information；Display module 3 passes through the related data information using display display patient.
The above is only the preferred embodiments of the present invention, and is not intended to limit the present invention in any form,
Any simple modification made to the above embodiment according to the technical essence of the invention, equivalent variations and modification, belong to
In the range of technical solution of the present invention.
Claims (5)
1. a kind of computer aided medicine data processing method, which is characterized in that the computer aided medicine data processing side
Method the following steps are included:
The first step carries out input system to the data information of patient, and computer passes through wireless signal transceiver for the information of patient
It is transmitted in Cloud Server store and share；
Second step carries out classified storage in reservoir to the data information of patient, is timely updated to the information of patient；
Third step, when data information of the clinician to patient is inquired, by the name or certificate that input patient
Patient information is verified, the information of patient is inquired, carries out relevant diagnoses and treatment；
The related data information of 4th step, patient is shown by display, is checked for clinician.
2. computer aided medicine data processing method as described in claim 1, which is characterized in that the relevant information of patient
Classified storage is carried out, using improved unbalanced data SVM algorithm, comprising the following steps:
Step 1, is arranged the most class sample point number MajorN=m*MinorN sampled in advance, and m is the quantitative ratio of the two；
Step 2 establishes the similar matrix S ∈ R based on Gaussian kernel first with most class sample setsn×n, n is original more
The number of several classes of samples；
Step 3 carries out clustering to most class samples using above-mentioned spectral clustering, generates cluster A1..., Ak, k=
MajorN；
Step 4 selects sample point representative in each cluster, wherein the samples selection number in each cluster
The size of average distance depending on sample and minority class sample point in the size of the cluster and the cluster, cluster is bigger,
The sample number of selection is more；More close from minority class sample, what is selected is fewer；
KDistil=K (xi, xi)+K(xl, xl)-2K(xi, xl)；
Wherein KsizeiTo cluster AiSize, IDistiTo cluster AiTo the average distance of minority class,It is each
The number of samples selected in cluster；
Step 5, for each cluster, before selectionIt is a from the smallest sample group of minority class sample point average distance
New most class training sample subsets are formed after conjunction；
Step 6, most class training sample subsets that sampling is obtained and whole minority class sample group cooperations are new training sample
This, be input in SVM algorithm and be trained study, wherein nuclear parameter is identical as the nuclear parameter of similar matrix S, i.e., spectral clustering with
Svm classifier algorithm all carries out in a space；
Step 7 is recognized according to the classification that the classification interface that training obtains carries out new samples.
3. computer aided medicine data processing method as described in claim 1, which is characterized in that the data information of patient
It timely updates, using improved data more new algorithm, detailed process is as follows:
Original transaction database is set as DB, transaction set T={ T1, T2..., Tn, candidate CD, frequent item set LD, former
Beginning transaction database size is | DB |；Newly-increased database is db, frequent item set Ld, newly-increased Database size is | db |；It updates
Database afterwards is S, S=DB ∪ db, frequent item set Ls, support sup；
Input: original transaction data set DB；Newly-increased transaction data set (TDS) db；
Output: updated data set S；Frequent item set LS；
Step 1 scans raw data set, obtains 1 item collection of candidate of DBFrequent 1 item collection is obtained according to support
It is established after sequence, FList is grouped according to grouping strategy, constructs the frequent pattern tree (fp tree) of DB, is carried out to frequent pattern tree (fp tree) frequent
The excavation of item collection obtains the frequent item set L of DBD；
Step 2 scans newly-increased data set, obtains 1 item collection of candidate of dbIn read step 1Merging obtains more
1 item collection of candidate of data set S after newFrequent 1 item collection of S is obtained according to supportFList ' is established after sequence,
It is grouped according to grouping strategy, constructs the frequent pattern tree (fp tree) of db；
Step 3, to frequent pattern tree (fp tree) carry out frequent item set excavation, while in read step 1 DB frequent item set LD, to generation
Each frequent mode k and LDIt compares: if k belongs to LD, then k is the frequent episode of original data set, by the support of k and in LDIt is right
The support counting addition answered can obtain support counting of the k in S, if minimum support of the tale greater than S counts, plus
Enter the part frequent item set L ' of S, and from LDIt is middle to delete this；If k is not belonging to LD, then k is the frequent episode of newly-increased data set, cannot
It whether frequently to determine in S, is added into the Candidate Set C of Ss；Differentiate LDIn remaining item, if its support counting is greater than S
Minimum support counts, then L ' is added；
Step 4 scans raw data set, the Candidate Set C of S in read step threeS, judge whether it is the frequent item set of S, it will be frequent
Item collection merges with part frequent item set L ', the frequent item set L of data set S after just being updatedS。
4. the computer aided medicine data processing that a kind of perform claim requires the 1 computer aided medicine data processing method
System, which is characterized in that the computer-assisted medical data processing system includes:
MIM message input module is connect with central processing module, carries out typing information to the related data information of patient；
Clinician's enquiry module, connect with central processing module, and according to the identity information of patient, clinician carries out patient's letter
The reading of breath carries out subsequent treatment；
Display module is connect with central processing module, by the related data information for showing patient using display；
Transmission module is exported, is connect with central processing module, by being passed the data information of patient using wireless signal transmitter
It is delivered in Cloud Server, carries out access reference；
Cloud Server is connect with central processing module, provides relevant case for clinician, and carry out to the information of patient
Storage is consulted for reference for clinicians；
Storage module is connect with central processing module, carries out classified storage to the relevant information of patient；
Information updating module is connect with central processing module, is timely updated to the data information of the patient in system；
Patient information authentication module, connect with central processing module, and clinician is by the name or certificate of input patient to disease
People's information is verified.
5. a kind of using the medical treatment auxiliary platform for calculating machine medical assistance data processing method described in claims 1 to 3 any one.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910043458.8A CN109887603A (en) | 2019-01-17 | 2019-01-17 | A kind of computer-assisted medical data processing system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910043458.8A CN109887603A (en) | 2019-01-17 | 2019-01-17 | A kind of computer-assisted medical data processing system and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109887603A true CN109887603A (en) | 2019-06-14 |
Family
ID=66926151
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910043458.8A Withdrawn CN109887603A (en) | 2019-01-17 | 2019-01-17 | A kind of computer-assisted medical data processing system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109887603A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110706778A (en) * | 2019-09-09 | 2020-01-17 | 武汉联析医疗技术有限公司 | Intelligent point location management system assisting in medical puncture |
| CN113345577A (en) * | 2021-06-18 | 2021-09-03 | 北京百度网讯科技有限公司 | Diagnosis and treatment auxiliary information generation method, model training method, device, equipment and storage medium |
-
2019
- 2019-01-17 CN CN201910043458.8A patent/CN109887603A/en not_active Withdrawn
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110706778A (en) * | 2019-09-09 | 2020-01-17 | 武汉联析医疗技术有限公司 | Intelligent point location management system assisting in medical puncture |
| CN110706778B (en) * | 2019-09-09 | 2022-08-09 | 武汉联析医疗技术有限公司 | Intelligent point location management system assisting in medical puncture |
| CN113345577A (en) * | 2021-06-18 | 2021-09-03 | 北京百度网讯科技有限公司 | Diagnosis and treatment auxiliary information generation method, model training method, device, equipment and storage medium |
| CN113345577B (en) * | 2021-06-18 | 2022-12-20 | 北京百度网讯科技有限公司 | Diagnosis and treatment auxiliary information generation method, model training method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kumar et al. | An integration of blockchain and AI for secure data sharing and detection of CT images for the hospitals | |
| CN110659692B (en) | Pathological image automatic labeling method based on reinforcement learning and deep neural network | |
| WO2020233254A1 (en) | Medical data analysis system integrating structured image data | |
| WO2021121126A1 (en) | Deep learning-based brain tumor image division method, device, apparatus, and medium | |
| WO2021051875A1 (en) | Cell classification method and apparatus, medium and electronic device | |
| WO2021233017A1 (en) | Image processing method and apparatus, and device and computer-readable storage medium | |
| Ye et al. | Glioma grading based on 3D multimodal convolutional neural network and privileged learning | |
| CN113724228A (en) | Tongue color and coating color identification method and device, computer equipment and storage medium | |
| CN110335668A (en) | Thyroid cancer cell pathological map auxiliary analysis method and system based on deep learning | |
| Zhang et al. | Urine sediment recognition method based on multi-view deep residual learning in microscopic image | |
| de Sousa Costa et al. | Classification of malignant and benign lung nodules using taxonomic diversity index and phylogenetic distance | |
| CN101699515B (en) | Multi-elite immune quantum clustering-based medical image segmenting system and multi-elite immune quantum clustering-based medical image segmenting method | |
| CN112132827A (en) | Pathological image processing method and device, electronic equipment and readable storage medium | |
| WO2024108663A1 (en) | Tumor survival prediction method and apparatus, electronic device and storage medium | |
| CN109887603A (en) | A kind of computer-assisted medical data processing system and method | |
| CN110400610A (en) | Small sample clinical data classification method and system based on multichannel random forest | |
| Na et al. | Retinal vascular segmentation using superpixel‐based line operator and its application to vascular topology estimation | |
| Elayaraja et al. | An efficient approach for detection and classification of cancer regions in cervical images using optimization based CNN classification approach | |
| Zhang et al. | Abnormal region detection in cervical smear images based on fully convolutional network | |
| Hao et al. | VP-Detector: A 3D multi-scale dense convolutional neural network for macromolecule localization and classification in cryo-electron tomograms | |
| CN114864075A (en) | Glioma grade analysis method and device based on pathological image | |
| Cheng et al. | Contrastive learning for echocardiographic view integration | |
| Meng et al. | Clinical applications of graph neural networks in computational histopathology: A review | |
| CN113705318B (en) | Image-based identification method, device, equipment and readable storage medium | |
| WO2021114626A1 (en) | Method for detecting quality of medical record data and related device |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20190614 |
|
| WW01 | Invention patent application withdrawn after publication |