CN101517602A - Methods for feature selection using classifier ensemble based genetic algorithms - Google Patents
Methods for feature selection using classifier ensemble based genetic algorithms Download PDFInfo
- Publication number
- CN101517602A CN101517602A CNA2007800347299A CN200780034729A CN101517602A CN 101517602 A CN101517602 A CN 101517602A CN A2007800347299 A CNA2007800347299 A CN A2007800347299A CN 200780034729 A CN200780034729 A CN 200780034729A CN 101517602 A CN101517602 A CN 101517602A
- Authority
- CN
- China
- Prior art keywords
- data
- genetic algorithm
- group
- sorters
- following composition
- 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
Images
Abstract
Methods for performing genetic algorithm-based feature selection are provided herein. In certain embodiments, the methods include steps of applying multiple data splitting patterns to a learning data set to build multiple classifiers to obtain at least one classification result; integrating the at least one classification result from the multiple classifiers to obtain an integrated accuracy result; and outputting the integrated accuracy result to a genetic algorithm as a fitness value for a candidate feature subset, in which genetic algorithm-based feature selection is performed.
Description
The sequence number that the application requires on September 22nd, 2006 to submit to is the rights and interests of 60/826,593 U.S. Provisional Application, and the full content of above-mentioned application is incorporated herein with way of reference.
Provide and used genetic algorithm to carry out the method for feature selecting.
Genetic algorithm (GA) is to be used for calculating to find out as search technique optimizing and a class evolution algorithmic of the solution of search problem.GA uses term and notion to develop to be subjected to the technology that biology inspires that develops, and comprises such as heredity, suddenlys change, selects and intersect such notion.
Feature selecting is also referred to as subclass and selects or Variables Selection, is a kind of method of using in machine learning.Before learning algorithm was applied to data set, selection can be from the subclass of this data set acquisition.It is infeasible on calculating that the process that use characteristic is selected, reason are to use all obtainable features of data centralization.Feature selecting also is used to make the problem of estimating and crossing fitness few generation of trying one's best when data set has the finite data sample that comprises big measure feature.
The typical fields of utilizing feature selecting is computer-aided diagnosis (CADx).CADx is that a kind of use machine learning techniques is predicted medical outcome, for example unknown pathology is categorized as pernicious or optimum method.For example, in computer tomography (CT) imaging of the lung that is used for pulmonary cancer diagnosis, these input feature vectors can comprise the result of the image processing algorithm that is applied to lung tubercle to be studied.The diagnosis accuracy that improves the CADx system is that this technology is successfully introduced committed step in clinical.
Owing to may need to calculate and retrieve a large amount of characteristics of image and Clinical symptoms for each pathology, owing to can not use all obtainable features of data centralization and when data set has the finite data sample that comprises big measure feature, have the problem of estimating, thereby feature selecting is an important step.Proved that adopting the feature selecting of GA and support vector machine (SVM) is to be used for computer aided detection (CAD; People such as Boroczky, IEEE Transaction on Biomedical Engineering, 10 (3), 504-551 page or leaf, 2006) efficient feature selection approach.
Although in having proved in a lot of fields based on the feature selecting of GA is successful, owing to have noise and medical data collection little and problem and deviation usually take place.This random division by GA inside causes, and described random division can generate deviation training dataset and test bias data set from learning data set.
Therefore, this paper provides the method that is used to carry out based on the feature selecting of genetic algorithm.Described method may further comprise the steps in one embodiment: a plurality of data are cut apart pattern be applied to learning data set to set up a plurality of sorters and then to obtain at least one classification results; Integration is from the accuracy result of described at least one classification results to obtain to integrate of described a plurality of sorters; And the accuracy result of described integration outputed to genetic algorithm as the fitness value that is used for candidate feature subset, wherein, carry out feature selecting based on genetic algorithm.
A related embodiment also comprises uses described genetic algorithm to obtain described candidate feature subset.
In a related embodiment, described a plurality of data are cut apart pattern described learning data set are divided into training data and test data.Learning data set is used to regulate the parameter of learning rules.Training dataset comprises that input vector (comprising obtainable feature) and answer vector (comprise known diagnosis, be malignant/benign), and use described training dataset so that adopt database training computer with case and known diagnosis with the supervised learning method.Test data set comprises the known embodiment for use in testing for the performance of the sorter of setting up on the training data.
In another related embodiment, described a plurality of sorters are selected from the group of at least one following composition: support vector machine, decision tree, linear discriminant analysis and neural network.
In another related embodiment, set up described a plurality of sorter and also comprise and use again Sampling techniques to obtain a plurality of training sets and a plurality of test set each from described learning data set.
In another related embodiment, set up described a plurality of sorter and also comprise a plurality of training sets of use.
In another embodiment, described method comprises that also combination predicts with the formation group from the classification results of described a plurality of sorters.
In a related embodiment, integrate at least one classification results and comprise that also calculating is selected from least one result of following group: mean value, weighted mean value, most ballot, the most ballots of weighting and intermediate value.
In another related embodiment, described method also comprises the employing genetic algorithm so that adopt described fitness value evaluate candidate feature subsets repeatedly, thereby generates new candidate feature subset, and obtains best final character subset.
In a related embodiment, described method is used for being selected from the medical imaging mode of the group of following at least one: CT, MRI, X ray and ultrasonic.
In another embodiment, described method is used for computer aided detection (CAD).In a related embodiment, described method is used for being selected from the CAD of disease of the group of at least one following composition: lung cancer, breast cancer, prostate cancer and colorectal cancer.
In another embodiment, described method is used for computer-aided diagnosis (CADx).In a related embodiment, described method is used for being selected from the CADx of disease of the group of at least one following composition: lung cancer, breast cancer, prostate cancer and colorectal cancer.
Method provided herein is incorporated in the evolution feature selection process categorizer integration method to improve the feature selecting based on GA.GA adopts cuts apart each character subset of consolidated forecast evaluation of result of pattern based on a plurality of data, rather than estimates single data and cut apart pattern.This is particularly useful for and may otherwise causes fitness value calculation noise data devious.
Fig. 1 shows data and cuts apart bar chart to the influence of classify accuracy;
Fig. 2 shows and sets up a plurality of sorters to analyze the process flow diagram of data set and the step that obtains optimal feature subset.
Feature selecting is used for determining that optimal feature subset is so that set up sorter.Use is based on the feature selection process of GA and SVM.Set up sorter based on optimal feature subset.
Sorter is used for the CAD and the CADx of various disease, for example is used for lung cancer and the cancer with other types of entity tumor.In the field of machine learning, sorter is used for the item with similar characteristics value is divided into groups.Possible sorter comprises SVM, decision tree, linear discriminant analysis and neural network.SVM is a linear classifier, and since its have the performance of outstanding relevant sorter and usually be used.Decision tree is a kind of forecast model, and it will be mapped as the conclusion about this desired value about one observation.Linear discriminant analysis is used to find out the linear combination of the feature of the object of distinguishing two or more classifications best or incident.The combination that obtains is as linear classifier or be used for carrying out dimensionality reduction before afterwards the classification.Neural network is a kind of non-linear statistical modeling tool, and it is used for pattern that the relation between the input and output is carried out modeling and/or found out data.
Provide the CADx system of high confidence level to diagnose (false positive and false negative are still less) to improve clinician's workflow fast and accurately by providing for the clinician.The CADx system can be as second proof-reader (reader) increasing the degree of confidence of clinician in its diagnosis, thereby make the unnecessary biopsy that tuberculosis is become (for example tubercle) significantly reduce, and make that significantly having reduced unnecessary treatment postpones.In addition, the CADx system can be so that carry out screening lung cancer to the asymptomatic patient, and reason is to diagnose fast and accurately.For example but the resolution that is not limited to the MSCT scanner of Philip Brilliance series and provides increased and allowed to observe trickleer structure, produced simultaneously image data amount for radiologist's interpretation obtains increasing.
In the CADx field based on machine learning, a modal problem is that training data has noise usually.Noise is especially serious when training dataset is enough not big.This has considerable influence to feature selecting validity.Cut apart and estimate the chromosome that each represents character subset because GA relies on random data, so the noise data inaccurate evaluation that provides character subset how to carry out.As a result, good character subset can be owing to its performance of cutting apart at " bad " random data is dropped.Whether this has influence on subsequently can successfully converge to optimal feature subset.
Fig. 1 shows the figure of employing from the experimental result of the data of 129 lung cancer cases.The data subset of selecting at random is used for training, that is, sets up the svm classifier device, and remaining data are used for test.This is called as data and cuts apart.Result among Fig. 1 shows when using different data to cut apart, classify accuracy, and promptly test accuracy is obviously different.
Method in the past supposes that typically noise component randomly draws from bias free (being average out to zero) normal distribution.Typically by the estimating noise deviation with deduct noise bias from fitness value and proofread and correct fitness value (people such as Miller, Evolutionary Computation 1996, can obtain at http://leitl.org/docs/ecj96.ps.gz).Fitness value is the objective measurement of the quality of separating.
Be not that all data in the real world all have unbiased distribution, perhaps deviation is difficult to estimate.In order to address these problems, The noise when during method provided herein uses the integrated GA of being reduced in of sorter to develop character subset being estimated.
Sorter be integrated in theory with experience on be proved to be than forming integrated any single sorter more accurate people such as (, Journal of Artificial Intelligence Research, 169-198 page or leaf, 1999) Opitz.Method provided herein is used following the variation: rely on Sampling techniques again and obtain to be used to set up the different training sets of a plurality of sorters and to use a plurality of character subsets to set up a plurality of sorters.To combine with the formation group from the classification results of a plurality of sorters and predict.
Be different from and set up a sorter (that is, using data to cut apart pattern) according to existing method with the performance of evaluating characteristic subclass, method provided herein is set up a plurality of sorters, be also referred to as integratedly, and integration is from the classification results of these sorters.In this case, on cutting apart, different pieces of information sets up several sorters.Each sorter will produce a decision-making, and for example pathology is pernicious or optimum.Integration method can be most ballots, that is, and and by the prediction of most member classifying devices selections.The integration method that substitutes comprises calculating mean value, weighted mean value or intermediate value (Kuncheva, L.I., IEEE Transactionson Pattern Analysis and Machine Intelligence, 24 (2), 281-286 page or leaf, 2002).Accuracy by the integrated acquisition of sorter is better than any single sorter.To return GA as the fitness value that is used for a specific characteristics subclass by the integrated definite integration accuracy of sorter.
Fig. 2 shows and is divided into two set, the data sample of set A (learning data set) and set B (giving over to the data set of final test).Set A experience data are cut apart, and the set A data are divided into training set and test set.Use a plurality of data and cut apart pattern to set up a plurality of sorters, i.e. SVM.Result from a plurality of sorters is integrated and estimates.Carry out the accuracy of classification on as the test set data of the part of raw data set.To return GA as the fitness value that is used for candidate feature subset as result from the classify accuracy of the integrated results of each sorter.Fitness value can comprise specificity and sensitivity.After integrated results was returned GA, GA determined which feature will be retained/abandon and suddenly change and interlace operation generation (one or more) new candidate feature subset by inner.Repeat the GA evolutionary process up to arriving end condition, determine optimal feature subset this moment.
Method provided herein can be used for some kinds of image modes, for example MRI, CT, X ray or ultrasonic.Method provided herein being applied to medical imaging mode, comprising the abnormality that is used for detecting and diagnosing human body, for example is the image mode of the data of electronic scanner collection from imaging system.Method and system provided herein can be used for the radiation work station, such as but not limited to Philip ExtendedBrilliance workstation, Philip Mx8000 and Philip Brilliance CT scan device series, or be integrated in the PACS system, such as but not limited to Stentor iSite system.Invention provided herein also is used for CAD and CADx.When being applied to CAD and CADx, invention provided herein is used to detect and diagnoses disease and other carcinous and non-cancerous lesion such as lung cancer, colorectal polyp, colorectal cancer, prostate cancer and breast cancer.
In addition will be obviously can under the situation of the spirit and scope that do not break away from claim and equivalent thereof, design other and other form of the present invention, and the embodiment of design except above-mentioned specific and one exemplary embodiment, so wish that scope of the present invention comprises these equivalents and instructions and claims and is intended to exemplary and is not to be construed as further restriction.The content of all lists of references that this paper quotes is incorporated herein with way of reference.
Claims (14)
1, a kind of method that is used to carry out based on the feature selecting of genetic algorithm, described method comprises:
A plurality of data are cut apart pattern be applied to learning data set setting up a plurality of sorters, and then obtain at least one classification results;
Integration is from the accuracy result of described at least one classification results to obtain to integrate of described a plurality of sorters; And
The accuracy result of described integration is outputed to genetic algorithm as the fitness value that is used for candidate feature subset, wherein, carry out feature selecting based on genetic algorithm.
2, method according to claim 1 also comprises and uses described genetic algorithm to obtain described candidate feature subset.
3, method according to claim 1, wherein, described a plurality of data are cut apart pattern described learning data set are divided into training data and test data.
4, method according to claim 1, wherein, described a plurality of sorters are selected from the group of at least one following composition: support vector machine, decision tree, linear discriminant analysis and neural network.
5, method according to claim 1 wherein, is set up described a plurality of sorter and is also comprised and use Sampling techniques to obtain a plurality of training sets and a plurality of test set each from described learning data set again.
6, method according to claim 1 wherein, is set up described a plurality of sorter and is also comprised a plurality of training sets of use.
7, method according to claim 1 comprises that also combination predicts with the formation group from the classification results of described a plurality of sorters.
8, method according to claim 1 wherein, is integrated at least one classification results and is also comprised at least one result who calculates the group that is selected from following composition: mean value, weighted mean value, most ballot, the most ballots of weighting and intermediate value.
9, method according to claim 1 also comprises and uses genetic algorithm to obtain best final character subset.
10, method according to claim 1, wherein, described method is used for being selected from the medical imaging mode of the group of at least one following composition: CT, MRI, X ray and ultrasonic.
11, method according to claim 1, wherein, described method is used for computer aided detection (CAD).
12, method according to claim 11, wherein, described method is used for being selected from the CAD of disease of the group of at least one following composition: lung cancer, breast cancer, prostate cancer and colorectal cancer.
13, method according to claim 1, wherein, described method is used for computer-aided diagnosis (CADx).
14, method according to claim 13, wherein, described method is used for being selected from the CADx of disease of the group of at least one following composition: lung cancer, breast cancer, prostate cancer and colorectal cancer.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US82659306P | 2006-09-22 | 2006-09-22 | |
| US60/826,593 | 2006-09-22 | ||
| US60/884,288 | 2007-01-10 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101517602A true CN101517602A (en) | 2009-08-26 |
Family
ID=41040569
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA2007800347299A Pending CN101517602A (en) | 2006-09-22 | 2007-09-17 | Methods for feature selection using classifier ensemble based genetic algorithms |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101517602A (en) |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102087721A (en) * | 2009-12-04 | 2011-06-08 | 索尼公司 | Information processing apparatus, observation value prediction method, and program |
| CN102708376A (en) * | 2012-04-19 | 2012-10-03 | 中国人民解放军总参谋部第六十三研究所 | Combined classifying device implementing system and method based on single-source information fusion |
| CN104866713A (en) * | 2015-05-12 | 2015-08-26 | 南京霁云信息科技有限公司 | Kawasaki disease and fever diagnosis system based on embedding of incremental local discrimination subspace |
| CN105550578A (en) * | 2015-12-10 | 2016-05-04 | 上海电机学院 | Network anomaly classification rule extracting method based on feature selection and decision tree |
| CN105651941A (en) * | 2015-12-31 | 2016-06-08 | 云南中烟工业有限责任公司 | Intelligent cigarette sensory evaluation system based on decomposition-aggregation strategy |
| CN105738908A (en) * | 2016-01-29 | 2016-07-06 | 宇龙计算机通信科技(深圳)有限公司 | Anti-collision early warning method, anti-collision early warning device, and earphone |
| CN106066936A (en) * | 2016-05-30 | 2016-11-02 | 北京千安哲信息技术有限公司 | A kind of method and system improving adverse effect precision of prediction |
| CN106709572A (en) * | 2015-11-16 | 2017-05-24 | 阿里巴巴集团控股有限公司 | Data processing method and equipment |
| CN106778005A (en) * | 2016-12-27 | 2017-05-31 | 中南民族大学 | Prostate cancer computer aided detection method and system based on multi-parameter MRI |
| CN107563997A (en) * | 2017-08-24 | 2018-01-09 | 京东方科技集团股份有限公司 | A kind of skin disease diagnostic system, construction method, diagnostic method and diagnostic device |
| CN109376798A (en) * | 2018-11-23 | 2019-02-22 | 东南大学 | A kind of classification method based on convolutional neural networks titanium dioxide lattice phase |
| CN109659026A (en) * | 2018-10-30 | 2019-04-19 | 平安科技（深圳）有限公司 | Disease kind analysis method and device, storage medium, the computer equipment of artificial intelligence |
| CN110574049A (en) * | 2017-05-19 | 2019-12-13 | 谷歌有限责任公司 | Multitask multimodal machine learning system |
| CN111066061A (en) * | 2017-09-11 | 2020-04-24 | 富士通株式会社 | Information processing apparatus, information processing method, and information processing program |
| CN112259221A (en) * | 2020-10-21 | 2021-01-22 | 北京大学第一医院 | Lung cancer diagnosis system based on multiple machine learning algorithms |
| CN112450982A (en) * | 2020-12-09 | 2021-03-09 | 上海市闵行区中心医院 | Qualitative and data scoring system for predicting prostate cancer through B-ultrasonic expression |
| CN114202524A (en) * | 2021-12-10 | 2022-03-18 | 中国人民解放军陆军特色医学中心 | Performance evaluation method and system of multi-modal medical image |
| WO2022141201A1 (en) * | 2020-12-30 | 2022-07-07 | 深圳高性能医疗器械国家研究院有限公司 | Breast cancer grading method based on dce-mri |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1670764A (en) * | 2004-03-19 | 2005-09-21 | 中国科学院计算技术研究所 | Genetic algorithm based human face sample generating method |
-
2007
- 2007-09-17 CN CNA2007800347299A patent/CN101517602A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1670764A (en) * | 2004-03-19 | 2005-09-21 | 中国科学院计算技术研究所 | Genetic algorithm based human face sample generating method |
Non-Patent Citations (1)
| Title |
|---|
| DAVID W.OPTIZ: "Feature Selection for Ensembles", 《AAAI-99 PROCEEDINGS》 * |
Cited By (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102087721B (en) * | 2009-12-04 | 2015-04-01 | 索尼公司 | Information processing apparatus, observation value prediction method |
| CN102087721A (en) * | 2009-12-04 | 2011-06-08 | 索尼公司 | Information processing apparatus, observation value prediction method, and program |
| CN102708376A (en) * | 2012-04-19 | 2012-10-03 | 中国人民解放军总参谋部第六十三研究所 | Combined classifying device implementing system and method based on single-source information fusion |
| CN102708376B (en) * | 2012-04-19 | 2015-10-21 | 中国人民解放军总参谋部第六十三研究所 | A kind of system and method realizing the assembled classifier merged based on single source information |
| CN104866713A (en) * | 2015-05-12 | 2015-08-26 | 南京霁云信息科技有限公司 | Kawasaki disease and fever diagnosis system based on embedding of incremental local discrimination subspace |
| CN104866713B (en) * | 2015-05-12 | 2018-02-13 | 南京霁云信息科技有限公司 | Locally differentiate the Kawasaki disease and fever diagnostic system of subspace insertion based on increment |
| CN106709572A (en) * | 2015-11-16 | 2017-05-24 | 阿里巴巴集团控股有限公司 | Data processing method and equipment |
| CN105550578A (en) * | 2015-12-10 | 2016-05-04 | 上海电机学院 | Network anomaly classification rule extracting method based on feature selection and decision tree |
| CN105651941A (en) * | 2015-12-31 | 2016-06-08 | 云南中烟工业有限责任公司 | Intelligent cigarette sensory evaluation system based on decomposition-aggregation strategy |
| CN105738908A (en) * | 2016-01-29 | 2016-07-06 | 宇龙计算机通信科技(深圳)有限公司 | Anti-collision early warning method, anti-collision early warning device, and earphone |
| CN105738908B (en) * | 2016-01-29 | 2018-08-24 | 宇龙计算机通信科技(深圳)有限公司 | A kind of anticollision method for early warning, device and earphone |
| CN106066936A (en) * | 2016-05-30 | 2016-11-02 | 北京千安哲信息技术有限公司 | A kind of method and system improving adverse effect precision of prediction |
| CN106778005B (en) * | 2016-12-27 | 2019-06-07 | 中南民族大学 | Prostate cancer computer-aided detection system based on multi-parameter MRI |
| CN106778005A (en) * | 2016-12-27 | 2017-05-31 | 中南民族大学 | Prostate cancer computer aided detection method and system based on multi-parameter MRI |
| CN110574049A (en) * | 2017-05-19 | 2019-12-13 | 谷歌有限责任公司 | Multitask multimodal machine learning system |
| CN110574049B (en) * | 2017-05-19 | 2023-05-16 | 谷歌有限责任公司 | Multi-task multi-modal machine learning system |
| CN107563997A (en) * | 2017-08-24 | 2018-01-09 | 京东方科技集团股份有限公司 | A kind of skin disease diagnostic system, construction method, diagnostic method and diagnostic device |
| CN107563997B (en) * | 2017-08-24 | 2020-06-02 | 京东方科技集团股份有限公司 | Skin disease diagnosis system, construction method, classification method and diagnosis device |
| CN111066061A (en) * | 2017-09-11 | 2020-04-24 | 富士通株式会社 | Information processing apparatus, information processing method, and information processing program |
| CN109659026A (en) * | 2018-10-30 | 2019-04-19 | 平安科技（深圳）有限公司 | Disease kind analysis method and device, storage medium, the computer equipment of artificial intelligence |
| CN109376798A (en) * | 2018-11-23 | 2019-02-22 | 东南大学 | A kind of classification method based on convolutional neural networks titanium dioxide lattice phase |
| CN109376798B (en) * | 2018-11-23 | 2021-09-24 | 东南大学 | Titanium dioxide lattice phase classification method based on convolutional neural network |
| CN112259221A (en) * | 2020-10-21 | 2021-01-22 | 北京大学第一医院 | Lung cancer diagnosis system based on multiple machine learning algorithms |
| CN112450982A (en) * | 2020-12-09 | 2021-03-09 | 上海市闵行区中心医院 | Qualitative and data scoring system for predicting prostate cancer through B-ultrasonic expression |
| CN112450982B (en) * | 2020-12-09 | 2023-03-28 | 上海市闵行区中心医院 | Qualitative and data scoring system for predicting prostate cancer through B-ultrasonic expression |
| WO2022141201A1 (en) * | 2020-12-30 | 2022-07-07 | 深圳高性能医疗器械国家研究院有限公司 | Breast cancer grading method based on dce-mri |
| CN114202524A (en) * | 2021-12-10 | 2022-03-18 | 中国人民解放军陆军特色医学中心 | Performance evaluation method and system of multi-modal medical image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101517602A (en) | Methods for feature selection using classifier ensemble based genetic algorithms | |
| EP2070024B1 (en) | Methods for feature selection using classifier ensemble based genetic algorithms | |
| Arunkumar et al. | Fully automatic model‐based segmentation and classification approach for MRI brain tumor using artificial neural networks | |
| Chen et al. | CytoBrain: cervical cancer screening system based on deep learning technology | |
| CN102165454B (en) | For improving the method for computer-aided diagnosis to the probabilistic robustness of image procossing | |
| Iannello et al. | Mitotic cells recognition in HEp-2 images | |
| CN107679368A (en) | PET/CT high dimensional feature level systems of selection based on genetic algorithm and varied precision rough set | |
| Chow et al. | Quantitative and qualitative analysis of 18 deep convolutional neural network (CNN) models with transfer learning to diagnose COVID-19 on Chest X-Ray (CXR) Images | |
| Koziarski et al. | DiagSet: a dataset for prostate cancer histopathological image classification | |
| Godbin et al. | Screening of COVID-19 based on GLCM features from CT images using machine learning classifiers | |
| CN117315379B (en) | Deep learning-oriented medical image classification model fairness evaluation method and device | |
| Asodekar et al. | Brain tumor classification using shape analysis of MRI images | |
| CN116864011A (en) | Colorectal cancer molecular marker identification method and system based on multiple sets of chemical data | |
| Sameki et al. | ICORD: Intelligent Collection of Redundant Data-A Dynamic System for Crowdsourcing Cell Segmentations Accurately and Efficiently. | |
| US20220044762A1 (en) | Methods of assessing breast cancer using machine learning systems | |
| de Oliveira Martins et al. | Classification of breast masses in mammogram images using Ripley’s K function and support vector machine | |
| Muhammad et al. | Brain Tumor Detection and Classification in Magnetic Resonance Imaging (MRI) using Region Growing, Fuzzy Symmetric Measure, and Artificial Neural Network Backpropagation | |
| Vijayadeep et al. | A hybrid feature extraction based optimized random forest learning model for brain stroke prediction | |
| Mishra et al. | Identification and detection of brain tumour using deep learning-based classification MRI applied using neural network and machine learning algorithm | |
| Tarekegn et al. | A cluster-genetic programming approach for detecting pulmonary tuberculosis | |
| Sahran et al. | Multi-level learning approach for prostate histopathology images classification | |
| Kumar et al. | Artificial intelligence bias in medical system designs: A systematic review | |
| Mehta et al. | Soft-computing based diagnostic tool for analyzing demyelination in magnetic resonance images | |
| Ghashghaei et al. | Grayscale Image Statistical Attributes Effectively Distinguish the Severity of Lung Abnormalities in CT Scan Slices of COVID-19 Patients | |
| Li et al. | Prediction of Short-Term Breast Cancer Risk with Fusion of CC-and MLO-Based Risk Models in Four-View Mammograms |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20090826 |