CN104252570A - Mass medical image data mining system and realization method thereof - Google Patents
Mass medical image data mining system and realization method thereof Download PDFInfo
- Publication number
- CN104252570A CN104252570A CN201310264654.0A CN201310264654A CN104252570A CN 104252570 A CN104252570 A CN 104252570A CN 201310264654 A CN201310264654 A CN 201310264654A CN 104252570 A CN104252570 A CN 104252570A
- Authority
- CN
- China
- Prior art keywords
- data
- medical image
- image data
- module
- mining
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Abstract
The invention provides a mass medical image data mining system and a realization method of the mass medical image data mining system. The system comprises a service application layer, an image mining layer and a data layer, wherein the image mining layer comprises a data preprocessing module, an image feature extraction module connected with the data preprocessing module and a data mining module connected with the image feature extraction module. The method comprises the following steps that a user makes requests at the service application layer; the data preprocessing module of the image mining layer obtains medical image data from the data layer and carries out preprocessing on the medical image data; the image feature extraction module analyzes the medical image data and obtains the integrated features of the medical image data; according to the integrated features, the data mining module carries out mining on the medical image data, and in addition, mining results are fed back to the service application layer. The technical scheme can be used for fast and accurately mining useful knowledge from mass medical image data sets.
Description
Technical field
The present invention relates to data processing field, particularly relate to a kind of massive medical image data digging system and its implementation.
Background technology
Along with the development of Medical Imaging Technology and computer technology, the imaging techniques such as X-ray imaging, computed tomography, magnetic resonance imaging, ultrasonic imaging and positron emission tomoscan and equipment play considerable effect in the clinical of medical institutions and R&D work.In order to realize sharing of medical image resource, domestic many medical institutions achieve information system management, picture archive and transmission system (Picture Archiving And Communication System, PACS) is used to gather medical image data, store, process and transmit.
The PACS system storage medical image data set of magnanimity.But, medical institutions are only limitted to the basic operations such as daily typing, inquiry and deletion for the operation of PACS system at present, be not easy to obtain valuable medical image resource accurately and efficiently, massive medical image resource cannot be fully utilized, cause the great wasting of resources.Therefore, how from massive medical image data, to obtain the information having potential value efficiently and accurately, for clinical diagnosis and Medical Research Work provide support, become current problem in the urgent need to address.
In the practical application of medical industry, there is following bottleneck in data mining technology.On the one hand, data mining technology is mainly used in structural data (such as, can carry out numeral or the symbol of logical expression realization by bivariate table structure), and medical image data is unstructured data mostly, the information contained is very complicated, cannot be directly used in data mining; On the other hand, the data processing method of existing business software generally adopts centralized processing, the data volume that can excavate and the efficiency of excavation depend on the performance of single computer platform, cannot excavate the information that medical worker needs economical, efficiently from the Medical imaging of explosive growth.
Summary of the invention
The problem that the present invention solves is to provide a kind of massive medical image data digging system of method and its implementation, can concentrate quickly and accurately excavate useful knowledge from massive medical image data.
In order to solve the problem, the invention provides a kind of massive medical image data digging system, comprise service application layer, the image tap layer be connected with described service application layer and the data Layer be connected with described service application layer and image tap layer, wherein, described image tap layer comprises data preprocessing module, the image feature extraction module be connected with described data preprocessing module and the data-mining module be connected with described image feature extraction module.
A kind of massive medical image data digging system described above, wherein, the image mining application interface that described service application layer comprises user interface and is connected with described user interface.
A kind of massive medical image data digging system described above, wherein, described image mining application interface comprises data prediction application programming interfaces, image feature extracts application programming interfaces and data mining application programming interfaces.
A kind of massive medical image data digging system described above, wherein, described image tap layer also comprises MapReduce Distributed Calculation module.
A kind of massive medical image data digging system described above, wherein, described data preprocessing module comprises format conversion submodule, normalization submodule, image denoising submodule and Iamge Segmentation submodule.
A kind of massive medical image data digging system described above, wherein, described image feature extraction module comprises Content Feature Extraction submodule, semantic feature extraction submodule and the integrated submodule of feature.
A kind of massive medical image data digging system described above, wherein, the data store and management module that described data Layer comprises data access interface and is connected with described data access interface.
In order to solve the problem, present invention also offers a kind of method for digging of massive medical image data, comprise the steps: that user files a request at described service application layer; The data preprocessing module of described image tap layer obtains medical image data from described data Layer, and carries out pre-service to described medical image data; Described image feature extraction module is analyzed described medical image data, and obtains the integration characteristic of described medical image data; According to described integration characteristic, described data-mining module excavates described medical image data, and gives described service application layer by the result feedback of described excavation.
Medical image data method for digging described above, wherein, by described MapReduce Distributed Calculation module, realizes the operation of described data preprocessing module, image feature extraction module and data-mining module.
Medical image data method for digging described above, wherein, described data-mining module is excavated described medical image data by weighted voting algorithm.
Medical image data method for digging described above, wherein, described Nearest Neighbor with Weighted Voting method is:
1) for the N kind method for digging in the arbitrary submodule in described data-mining module arranges weight, N >=1;
2) described N kind method for digging is used to excavate described medical image data respectively, and the Result of record each method for digging described;
3) calculate the weighted results of the Result of each method for digging described, and according to described weighted results, decision-making is carried out to described medical image data.
Medical image data method for digging described above, wherein, according to described Result, adjusts the weight of described N kind method for digging.
Compared with prior art, the present invention introduces image feature extraction module, the integration characteristic of medical image can be extracted, fully utilize the various features of medical image, have comprehensive, and avoid single features cannot the complete defect that must describe medical image content, more fully, accurately can express the attribute of picture material, improve the accuracy of medical image Result;
Further, introduce several data mining algorithm in data-mining module, adopt weighted voting algorithm to make a policy, take full advantage of the complementarity between different pieces of information mining algorithm, improve the accuracy of data mining results, and the weight of algorithms of different can also be adjusted according to the result of decision in real time;
Further, based on MapReduce Distributed Calculation module, the cluster that only need be made up of server that is cheap, low side, the size of cluster can be selected according to the actual loading of user, while guarantee medical image data digging system high efficiency, also for user has saved cost;
Further, image tap layer is decoupled into the module of three independent operatings, and each module is made up of multiple completely independently submodule again, has good expandability, according to the data mining demand of user, can add new functional module quickly and easily.
Accompanying drawing explanation
Figure 1 shows that the structural representation of a kind of massive medical image data of embodiment of the present invention digging system;
Figure 2 shows that the schematic flow sheet of a kind of massive medical image data of embodiment of the present invention method for digging;
Figure 3 shows that the schematic flow sheet of embodiment of the present invention weighted voting algorithm.
Embodiment
Set forth a lot of detail in the following description so that fully understand the present invention.But the present invention can be much different from alternate manner described here to implement, those skilled in the art can when without prejudice to doing similar popularization when intension of the present invention, therefore the present invention is by the restriction of following public concrete enforcement.
Secondly, the present invention utilizes schematic diagram to be described in detail, and when describing the embodiment of the present invention in detail, for ease of illustrating, described schematic diagram is example, and it should not limit the scope of protection of the invention at this.
Below in conjunction with drawings and Examples, a kind of massive medical image data digging system of the present invention and its implementation are described in detail.The massive medical image data digging system of the embodiment of the present invention as shown in Figure 1, described data digging system comprises service application layer 1, the image tap layer 2 be connected with described service application layer 1 and the data Layer 3 be connected with described service application layer 1 and image tap layer 2.Wherein, the image mining application interface that described service application layer 1 comprises user interface 11 and is connected with described user interface 11, pass through application programming interfaces, user can call each submodule of image tap layer 2, and described image mining application interface comprises data prediction application programming interfaces 12, image feature extracts application programming interfaces 13 and data mining application programming interfaces 14; Described image tap layer 2 is core layers of described medical image data digging system, comprise data preprocessing module 21, the image feature extraction module 22 be connected with described data preprocessing module and the data-mining module 23 be connected with described image feature extraction module 22, also comprise MapReduce Distributed Calculation module 24; The data store and management module 32 that described data Layer 3 comprises data access interface 31 and is connected with described data access interface 31, wherein, described data access interface 31 comprises medical image data access interface and expertise data access interface.
It should be noted that, three module independent operatings of image tap layer, each module is used for the different flow processs in the excavation of operational medicine image, and each module is made up of multiple completely independently submodule, and submodule is for implementing different data processing algorithms.Wherein, described data prediction mould 21 pieces comprises format conversion submodule, normalization submodule, image denoising submodule and Iamge Segmentation submodule; Described image feature extraction module 22 comprises Content Feature Extraction submodule, semantic feature extraction submodule and the integrated submodule of feature.Data-mining module 23 comprises multiple submodule, completely independent between submodule, each submodule is for realizing a kind of data mining capability, such as, classification submodule is used for classifying to medical image, cluster analysis submodule is used for carrying out cluster to medical image, and association analysis submodule is for finding the correlation rule of medical image inside.
As shown in Figure 2, first, perform step S201, user's method that medical image data digging system excavates described medical image data files a request at described service application layer.Particularly, in the present embodiment, the task that user selects and submit to medical image data to excavate by the user interface in described service application layer, the situation that namely in mining data storehouse, all patients suffer from breast cancer.
Then, perform step S202, the data preprocessing module of described image tap layer obtains medical image data from described data Layer, and carries out pre-service to described medical image data.Particularly, data processing module obtains medical image data from the data store and management module data Layer.Described data store and management module by the patient information data of medical image data access interface loading structure from medical image data source (database of the PACS system of medical institutions, region medical image data center and other storing medical image) and non-structured medical image data, or obtains non-structured medical field knowledge (diagnostic imaging as lung cancer, breast cancer is regular) by expertise data access interface from medical expert knowledge library.Data store and management module is except providing data store and management function, and the user feedback data that the intermediate data that also can produce image tap layer and result data, service application layer produce carries out store and management.In the present embodiment, data preprocessing module obtains the breast molybdenum target X ray image of all patients in the medical image data center of region from data Layer.
Data preprocessing module carries out pre-service to described medical image data, thus eliminates the noise in data, improves the quality of data, promotes the accuracy and efficiency of image processing and data mining.Data preprocessing module comprises multiple submodule, and as format conversion submodule, normalization submodule, image denoising submodule and Iamge Segmentation submodule, each submodule works alone, and realizes different data prediction functions.Format conversion submodule can be converted to unified form medical image, so that batch processing medical image data, because the medical image got from regional data center may have multiple format, as BMP, JPG, DICOM etc.Normalization submodule can be normalized the yardstick of medical image, color, to eliminate the impact of dimension on the image feature of image feature, because different by the size of the medical image of different PACS system acquisition, color, when extracting image feature, characterizing magnitudes difference can be very large, thus impact is based on the data mining results of image feature.Image denoising submodule can carry out denoising to medical image, level and smooth and enhancing processes, improve the quality of medical image, this is that medical image owing to getting from PACS system generally comes from X-ray production apparatus, CT machine, the digital medical image equipment such as NMR imaging instrument, in imaging process (digitized process), working condition due to imageing sensor is subject to imaging circumstances condition, the impact of the factors such as each element quality of sensor self, the medical image generated is with noise, if noise is not removed, extraction and the data mining results of image feature will be affected.Iamge Segmentation submodule is used for splitting area-of-interest in medical image, such as, user is loaded with 10000 breast molybdenum target X ray images from data source, wish to excavate the Local Features of lump in breast, judge whether patient suffers from malignant breast carcinomas, so, user can select the image partition methods such as threshold method, region-growing method, cluster analysis by Iamge Segmentation submodule, lump in breast is split from breast molybdenum target X ray image, to reduce the redundancy of data set, improve digging efficiency.The algorithm of all data prediction is all realized by MapReduce Distributed Calculation module above.The cluster that the support of MapReduce Distributed Calculation module is made up of a large amount of low-end server, by mapping (map) function, the Processing tasks of mass data is distributed on each node of cluster automatically, executes the task each nodal parallel, and produce the output of multiple intermediate result; Collect intermediate result by reduction (reduce) function again to export, and union operation is carried out to intermediate result output set.Such as, in the present embodiment, denoising is carried out to all X ray images, then all X ray images are divided into ten subsets, pass through mapping function, the denoising task of each subset is automatically distributed on each node of cluster and (namely distributes to each computing machine), each nodal parallel ground carries out denoising to the image be assigned to, and produces multiple intermediate result output (result that namely in image denoising process, each step produces); Again by reduction function, the denoising result produced by all nodes carries out collecting, union operation, namely completes the denoising to all X ray images.
It should be noted that, user can check the situation with control data pretreatment module by user interface, also described data preprocessing module be can call by the data prediction application programming interfaces in service application layer, cleaning and compression etc. to data realized.And the submodule in data preprocessing module can add according to actual needs.
Then, perform step S203, described image feature extraction module is analyzed described medical image data, and obtains the integration characteristic of described medical image data.Image feature extraction module is used for extracting the feature of medical image, and these features can be the identifiable physical features of human vision in medical image, also can be medical image to be thought to some parameter of definition.Particularly, in the present embodiment, image feature extraction module comprises Content Feature Extraction submodule, semantic feature extraction submodule and the integrated submodule of feature, Content Feature Extraction submodule and semantic feature extraction submodule are used for extracting the various features of medical image respectively, and the various features extracted are gathered by the integrated submodule of feature.Content Feature Extraction submodule is used for by multiple image analysis technology, extracts the physical features such as the color of medical image, texture and shape.Such as, the color characteristic of medical image is extracted by analyzing the average of grey level histogram, variance, degree of tilt and steepness; The textural characteristics of medical image is extracted by analyzing the contrast of gray level co-occurrence matrixes, the degree of correlation and entropy; The shape facility that son extracts medical image is described by moment invariants, Fourier.Semantic feature extraction submodule, for extracting the semantic feature of medical image, is mainly divided into above two kinds of situations.One is medical image with text label (keyword added as doctor, expert and title), can using text label directly as the semantic feature of medical image under this situation; Two be medical image without text label, can set up semantic model according to machine learning method under this situation, is automatically medical image generative semantics feature.The integrated submodule of feature is used for the parameter set in the image mining application interface of service application layer according to user, the multiple image feature that integrated user specifies.Such as, user have selected color, texture and shape three kinds of features in application programming interfaces, the integrated submodule of feature can carry out standardization respectively to three kinds of features, and pass through data integrating method, as principal component analysis (PCA) (principal component analysis, PCA), the integration characteristic of medical image is extracted.In the present embodiment, extract the color of all X ray images, texture and shape by content characteristic submodule, and by the integrated submodule of feature, these features are gathered, extract the integration characteristic of X ray image.
It should be noted that, the algorithm that above all image features extract all is realized by MapReduce Distributed Calculation module, and implementation procedure is as above-mentioned process of image being carried out to denoising.And user can extract application programming interfaces by image feature and call image feature extraction submodule.
Then, perform step S204, according to described integration characteristic, described data-mining module excavates described medical image data, and gives described service application layer by the result feedback of described excavation.Wherein, data-mining module comprises multiple submodule, and completely independent between submodule, each submodule is for realizing a kind of data mining capability.Such as, classification submodule is used for classifying to medical image, and cluster analysis submodule is used for carrying out cluster to medical image, and association analysis submodule is for finding the correlation rule of medical image inside.Each data mining submodule, all correspond to multiple implementation method.Such as, submodule of classifying includes decision tree, naive Bayesian, neural network, support vector machine and rough set Lung biopsy.
Data-mining module is excavated described medical image data by weighted voting algorithm, described weighted voting algorithm as shown in Figure 3, first, perform step S301, for the N kind method for digging in the arbitrary submodule in described data-mining module arranges weight, weight and be 1, N >=1.Particularly, the method during user selects for making a policy submodule by the image mining application interface of service application layer.In the present embodiment, the medical image (X ray image) extracting integration characteristic is classified, user have selected decision tree, naive Bayesian, neural network, support vector machine and rough set five kinds of sorting techniques, so, when medical image is classified, the result considering Lung biopsy made a policy, the initial weight of above-mentioned Lung biopsy is set to identical value, is 0.2.
Then, perform step S302, use described N kind method for digging to excavate described medical image data respectively, and the Result of record each method for digging described.Particularly, in the present embodiment, five kinds of sorting techniques in step S301 are used to excavate medical image respectively, and recording the Result of each sorting technique, the Result of decision tree, naive Bayesian, neural network, support vector machine and rough set is respectively 1,0,1,1,1.Wherein, 0 represents that patient image examination result is negative, and 1 represents that the image check result of patient is positive.
Then, perform step S303, calculate the weighted results of the Result of each method for digging described, and according to described weighted results, decision-making is carried out to described medical image data.Particularly, in the present embodiment, calculate the weighting classification result of Lung biopsy, obtaining this result is 0.8, compare known with appointment threshold value (specifying threshold value v=0.75), weighting classification result is greater than appointment threshold value, then carry out decision-making to described medical image, namely the medical imaging result of patient is positive, and namely in database, all patients suffer from breast cancer the excavation of situation.
It should be noted that, according to described Result, the weight of described N kind method for digging can be adjusted.Particularly, according to the performance of above-mentioned five kinds of sorting algorithms, the weight of its correspondence is adjusted in real time.Such as, from step S302 and step S303, Naive Bayes Classifier has made wrong judgement, therefore, the weight of Naive Bayes Classifier declines 0.02, and all the other in four sorter made correct judgement, then the weight of often kind of sorter increases by 0.005 respectively, and adjusted weight is applied in mining process next time, make mining process can mate medical image data set better, improve the accuracy of medical image Result further.
Although the present invention with preferred embodiment openly as above; but it is not for limiting the present invention; any those skilled in the art without departing from the spirit and scope of the present invention; the Method and Technology content of above-mentioned announcement can be utilized to make possible variation and amendment to technical solution of the present invention; therefore; every content not departing from technical solution of the present invention; the any simple modification done above embodiment according to technical spirit of the present invention, equivalent variations and modification, all belong to the protection domain of technical solution of the present invention.
Claims (12)
1. a massive medical image data digging system, it is characterized in that, comprise service application layer, the image tap layer be connected with described service application layer and the data Layer be connected with described service application layer and image tap layer, wherein, described image tap layer comprises data preprocessing module, the image feature extraction module be connected with described data preprocessing module and the data-mining module be connected with described image feature extraction module.
2. a kind of massive medical image data digging system as claimed in claim 1, is characterized in that, the image mining application interface that described service application layer comprises user interface and is connected with described user interface.
3. a kind of massive medical image data digging system as claimed in claim 2, is characterized in that, described image mining application interface comprises data prediction application programming interfaces, image feature extracts application programming interfaces and data mining application programming interfaces.
4. a kind of massive medical image data digging system as claimed in claim 1, it is characterized in that, described image tap layer also comprises MapReduce Distributed Calculation module.
5. a kind of massive medical image data digging system as claimed in claim 1, it is characterized in that, described data preprocessing module comprises format conversion submodule, normalization submodule, image denoising submodule and Iamge Segmentation submodule.
6. a kind of massive medical image data digging system as claimed in claim 1, it is characterized in that, described image feature extraction module comprises Content Feature Extraction submodule, semantic feature extraction submodule and the integrated submodule of feature.
7. a kind of massive medical image data digging system as claimed in claim 1, is characterized in that, the data store and management module that described data Layer comprises data access interface and is connected with described data access interface.
8. the method for digging of a kind of massive medical image data as claimed in claim 1, is characterized in that, comprise the steps:
User files a request at described service application layer; The data preprocessing module of described image tap layer obtains medical image data from described data Layer, and carries out pre-service to described medical image data; Described image feature extraction module is analyzed described medical image data, and obtains the integration characteristic of described medical image data; According to described integration characteristic, described data-mining module excavates described medical image data, and gives described service application layer by the result feedback of described excavation.
9. medical image data method for digging as claimed in claim 8, is characterized in that, by described MapReduce Distributed Calculation module, realize the operation of described data preprocessing module, image feature extraction module and data-mining module.
10. medical image data method for digging as claimed in claim 8, it is characterized in that, described data-mining module is excavated described medical image data by weighted voting algorithm.
11. medical image data method for digging as claimed in claim 10, it is characterized in that, described Nearest Neighbor with Weighted Voting method is:
1) for the N kind method for digging in the arbitrary submodule in described data-mining module arranges weight, N >=1;
2) described N kind method for digging is used to excavate described medical image data respectively, and the Result of record each method for digging described;
3) calculate the weighted results of the Result of each method for digging described, and according to described weighted results, decision-making is carried out to described medical image data.
12. medical image data method for digging as claimed in claim 11, is characterized in that, according to described Result, adjust the weight of described N kind method for digging.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310264654.0A CN104252570A (en) | 2013-06-28 | 2013-06-28 | Mass medical image data mining system and realization method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310264654.0A CN104252570A (en) | 2013-06-28 | 2013-06-28 | Mass medical image data mining system and realization method thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104252570A true CN104252570A (en) | 2014-12-31 |
Family
ID=52187458
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310264654.0A Pending CN104252570A (en) | 2013-06-28 | 2013-06-28 | Mass medical image data mining system and realization method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104252570A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105184074A (en) * | 2015-09-01 | 2015-12-23 | 哈尔滨工程大学 | Multi-modal medical image data model based medical data extraction and parallel loading method |
| CN105894088A (en) * | 2016-03-25 | 2016-08-24 | 苏州赫博特医疗信息科技有限公司 | Medical information extraction system and method based on depth learning and distributed semantic features |
| CN107808687A (en) * | 2016-09-08 | 2018-03-16 | 京东方科技集团股份有限公司 | Medical data acquisition method, processing method, cluster processing system and method |
| WO2018132997A1 (en) * | 2017-01-19 | 2018-07-26 | 中国科学院深圳先进技术研究院 | Radiomics-based survival prediction method and device |
| CN108416360A (en) * | 2018-01-16 | 2018-08-17 | 华南理工大学 | Cancer diagnosis system and method based on breast molybdenum target calcification feature |
| CN108665456A (en) * | 2018-05-15 | 2018-10-16 | 广州尚医网信息技术有限公司 | The method and system that breast ultrasound focal area based on artificial intelligence marks in real time |
| CN109711459A (en) * | 2018-12-24 | 2019-05-03 | 广东德诚科教有限公司 | User individual action estimation method, apparatus, computer equipment and storage medium |
| CN110781659A (en) * | 2018-07-11 | 2020-02-11 | 株式会社Ntt都科摩 | Text processing method and text processing device based on neural network |
| CN111179642A (en) * | 2020-02-17 | 2020-05-19 | 山西医科大学第一医院 | Digital medical image teaching and scientific research system |
| CN112349369A (en) * | 2020-11-27 | 2021-02-09 | 广州瀚信通信科技股份有限公司 | Medical image big data intelligent analysis method, system and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7266473B1 (en) * | 2006-02-15 | 2007-09-04 | Agilent Technologies, Inc. | Fast microarray expression data analysis method for network exploration |
| US20070271223A1 (en) * | 2006-05-18 | 2007-11-22 | John Xiaoming Zhang | Method and implementation of reliable consensus feature selection in biomedical discovery |
| CN101551854A (en) * | 2009-05-13 | 2009-10-07 | 西安电子科技大学 | A processing system of unbalanced medical image and processing method thereof |
| CN101599088A (en) * | 2008-11-18 | 2009-12-09 | 北京美智医疗科技有限公司 | The mining multi-dimensional data system and method for medical information system |
| US7724961B2 (en) * | 2006-09-08 | 2010-05-25 | Mitsubishi Electric Research Laboratories, Inc. | Method for classifying data using an analytic manifold |
-
2013
- 2013-06-28 CN CN201310264654.0A patent/CN104252570A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7266473B1 (en) * | 2006-02-15 | 2007-09-04 | Agilent Technologies, Inc. | Fast microarray expression data analysis method for network exploration |
| US20070271223A1 (en) * | 2006-05-18 | 2007-11-22 | John Xiaoming Zhang | Method and implementation of reliable consensus feature selection in biomedical discovery |
| US7724961B2 (en) * | 2006-09-08 | 2010-05-25 | Mitsubishi Electric Research Laboratories, Inc. | Method for classifying data using an analytic manifold |
| CN101599088A (en) * | 2008-11-18 | 2009-12-09 | 北京美智医疗科技有限公司 | The mining multi-dimensional data system and method for medical information system |
| CN101551854A (en) * | 2009-05-13 | 2009-10-07 | 西安电子科技大学 | A processing system of unbalanced medical image and processing method thereof |
Non-Patent Citations (3)
| Title |
|---|
| 张春芬: "《多分类器组合及其在医学图像分类中的应用》", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 王娟: "《决策树分类技术在疾病辅助诊断系统中的研究与应用》", 《中国优秀硕士学位论文全文数据库信息科技辑 》 * |
| 王恒: "《基于PMML的医学图像挖掘系统的设计与实现》", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105184074A (en) * | 2015-09-01 | 2015-12-23 | 哈尔滨工程大学 | Multi-modal medical image data model based medical data extraction and parallel loading method |
| CN105184074B (en) * | 2015-09-01 | 2018-10-26 | 哈尔滨工程大学 | A kind of medical data extraction and loaded in parallel method based on Multimodal medical image data model |
| CN105894088A (en) * | 2016-03-25 | 2016-08-24 | 苏州赫博特医疗信息科技有限公司 | Medical information extraction system and method based on depth learning and distributed semantic features |
| CN105894088B (en) * | 2016-03-25 | 2018-06-29 | 苏州赫博特医疗信息科技有限公司 | Based on deep learning and distributed semantic feature medical information extraction system and method |
| CN107808687B (en) * | 2016-09-08 | 2021-01-29 | 京东方科技集团股份有限公司 | Medical data acquisition method, processing method, cluster processing system and method |
| CN107808687A (en) * | 2016-09-08 | 2018-03-16 | 京东方科技集团股份有限公司 | Medical data acquisition method, processing method, cluster processing system and method |
| US11664121B2 (en) | 2016-09-08 | 2023-05-30 | Boe Technology Group Co., Ltd. | Medical data processing method, cluster processing system and method thereof |
| US11062803B2 (en) | 2016-09-08 | 2021-07-13 | Boe Technology Group Co., Ltd. | Medical data processing method, cluster processing system and method thereof |
| WO2018132997A1 (en) * | 2017-01-19 | 2018-07-26 | 中国科学院深圳先进技术研究院 | Radiomics-based survival prediction method and device |
| CN108416360B (en) * | 2018-01-16 | 2022-03-29 | 华南理工大学 | Cancer diagnosis system and method based on breast molybdenum target calcification features |
| CN108416360A (en) * | 2018-01-16 | 2018-08-17 | 华南理工大学 | Cancer diagnosis system and method based on breast molybdenum target calcification feature |
| CN108665456A (en) * | 2018-05-15 | 2018-10-16 | 广州尚医网信息技术有限公司 | The method and system that breast ultrasound focal area based on artificial intelligence marks in real time |
| CN110781659A (en) * | 2018-07-11 | 2020-02-11 | 株式会社Ntt都科摩 | Text processing method and text processing device based on neural network |
| CN109711459B (en) * | 2018-12-24 | 2019-11-15 | 广东德诚科教有限公司 | User individual action estimation method, apparatus, computer equipment and storage medium |
| CN109711459A (en) * | 2018-12-24 | 2019-05-03 | 广东德诚科教有限公司 | User individual action estimation method, apparatus, computer equipment and storage medium |
| CN111179642A (en) * | 2020-02-17 | 2020-05-19 | 山西医科大学第一医院 | Digital medical image teaching and scientific research system |
| CN112349369A (en) * | 2020-11-27 | 2021-02-09 | 广州瀚信通信科技股份有限公司 | Medical image big data intelligent analysis method, system and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104252570A (en) | Mass medical image data mining system and realization method thereof | |
| Mittal et al. | A comprehensive survey of image segmentation: clustering methods, performance parameters, and benchmark datasets | |
| US20210110196A1 (en) | Deep Learning Network for Salient Region Identification in Images | |
| AU2020349082B2 (en) | System to collect and identify skin conditions from images and expert knowledge | |
| Garcia Seco De Herrera et al. | Overview of the ImageCLEF 2015 medical classification task | |
| CN107274402A (en) | A kind of Lung neoplasm automatic testing method and system based on chest CT image | |
| CN112534424A (en) | Neural network based content distribution in online systems | |
| CN111242948B (en) | Image processing method, image processing device, model training method, model training device, image processing equipment and storage medium | |
| US20220269648A1 (en) | System and method for providing correlated content organization in an enterprise content management system based on a training set | |
| Saraswathi et al. | Survey on image segmentation via clustering | |
| CN113313170A (en) | Full-time global training big data platform based on artificial intelligence | |
| Faruqui et al. | Healthcare As a Service (HAAS): CNN-based cloud computing model for ubiquitous access to lung cancer diagnosis | |
| WO2021041342A1 (en) | Semantic image retrieval for whole slide images | |
| Rajeswari et al. | Gradient contouring and texture modelling based cad system for improved tb classification | |
| Nataraja et al. | Machine learning algorithms for heterogeneous data: a comparative study | |
| KR20200124887A (en) | Method and Apparatus for Creating Labeling Model with Data Programming | |
| Haripriya et al. | A survey paper on data mining techniques and challenges in distributed DICOM | |
| Majji et al. | Smart iot in breast cancer detection using optimal deep learning | |
| Megalooikonomou et al. | Medical data fusion for telemedicine | |
| Kowshika et al. | Enhancing the Automated Diagnosis System of Soft Tissue Tumors with Machine Learning Techniques | |
| CN110993091B (en) | Generating vectors from data | |
| EP4208873A1 (en) | Systems and methods for automatically managing image data | |
| Mohanty et al. | RETRACTED ARTICLE: Mass classification method in mammograms using correlated association rule mining | |
| Zhan et al. | Graph entropy-based clustering algorithm in medical brain image database | |
| CN116701962B (en) | Edge data processing method, device, computing equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20141231 |