Use of machine learning for classification of magneto cardiograms
IPC분류정보
국가/구분
United States(US) Patent
등록
국제특허분류(IPC7판)
A61B-005/04
A61B-005/00
G06K-009/00
G06N-003/08
G06F-019/00
출원번호
US-0288697
(2014-05-28)
등록번호
US-9173614
(2015-11-03)
발명자
/ 주소
Sternickel, Karsten
Szymanski, Boleslaw
Embrechts, Mark
출원인 / 주소
CardioMag Imaging, Inc.
대리인 / 주소
Yablon, Jay R.
인용정보
피인용 횟수 :
1인용 특허 :
22
초록▼
The use of machine learning for pattern recognition in magnetocardiography (MCG) that measures magnetic fields emitted by the electrophysiological activity of the heart is disclosed herein. Direct kernel methods are used to separate abnormal MCG heart patterns from normal ones. For unsupervised lear
The use of machine learning for pattern recognition in magnetocardiography (MCG) that measures magnetic fields emitted by the electrophysiological activity of the heart is disclosed herein. Direct kernel methods are used to separate abnormal MCG heart patterns from normal ones. For unsupervised learning, Direct Kernel based Self-Organizing Maps are introduced. For supervised learning Direct Kernel Partial Least Squares and (Direct) Kernel Ridge Regression are used. These results are then compared with classical Support Vector Machines and Kernel Partial Least Squares. The hyper-parameters for these methods are tuned on a validation subset of the training data before testing. Also investigated is the most effective pre-processing, using local, vertical, horizontal and two-dimensional (global) Mahanalobis scaling, wavelet transforms, and variable selection by filtering.
대표청구항▼
1. A method for automating the identification of meaningful features and the formulation of expert rules for classifying magnetocardiography data, comprising: applying a direct kernel transform to sensed data acquired from sensors sensing magnetic fields generated by a patient's heart activity, resu
1. A method for automating the identification of meaningful features and the formulation of expert rules for classifying magnetocardiography data, comprising: applying a direct kernel transform to sensed data acquired from sensors sensing magnetic fields generated by a patient's heart activity, resulting in transformed data; andidentifying said meaningful features and formulating said expert rules from said transformed data, using machine learning. 2. The method of claim 1, said kernel transform satisfying Mercer conditions. 3. The method of claim 1, said kernel transform comprising a radial basis function. 4. The method of claim 1, said applying a kernel transform comprising: assigning said transformed data to a first hidden layer of a neural network;applying training data descriptors as weights of said first hidden layer of said neural network; andcalculating weights of a second hidden layer of said neural network numerically. 5. The method of claim 1, further comprising: classifying said transformed data using direct kernel partial least square (DK-PLS) machine learning. 6. The method of claim 1, further comprising transforming said sensed data into wavelet domain data prior to applying said kernel transform by: applying a Daubechies wavelet transform to said sensed data. 7. The method of claim 6, further comprising: selecting features from said wavelet domain data which improve said classification of magnetocardiography data. 8. The method of claim 1, further comprising: normalizing said sensed data. 9. The method of claim 8, said normalizing said sensed data comprising: Mahalanobis scaling said sensed data. 10. The method of claim 1, further comprising: centering a kernel of said kernel transform. 11. An apparatus for automating the identification of meaningful features and the formulation of expert rules for classifying magnetocardiography data, comprising computerized storage, processing and programming for: applying a direct kernel transform to sensed data acquired from sensors sensing magnetic fields generated by a patient's heart activity, resulting in transformed data; andidentifying said meaningful features and formulating said expert rules from said transformed data, using machine learning. 12. The apparatus of claim 11, wherein kernel transform satisfies Mercer conditions. 13. The apparatus of claim 11, said kernel transform comprising a radial basis function. 14. The apparatus of claim 11, said computerized storage, processing and programming for applying a kernel transform further comprising computerized storage, processing and programming for: assigning said transformed data to a first hidden layer of a neural network;applying training data descriptors as weights of said first hidden layer of said neural network; andcalculating weights of a second hidden layer of said neural network numerically. 15. The apparatus of claim 11, further comprising computerized storage, processing and programming for: classifying said transformed data using direct kernel partial least square (DK-PLS) machine learning. 16. The apparatus of claim 11, further comprising using said computerized storage, processing and programming for transforming said sensed data into wavelet domain data prior to applying said kernel transform by—: applying a Daubechies wavelet transform to said sensed data. 17. The apparatus of claim 16, further comprising using said computerized storage, processing and programming for: selecting features from said wavelet domain data which improve said classification of magnetocardiography data. 18. The apparatus of claim 11, further comprising computerized storage, processing and programming for: normalizing said sensed data. 19. The apparatus of claim 18, said computerized storage, processing and programming for normalizing said sensed data comprising computerized storage, processing and programming for: Mahalanobis scaling said sensed data. 20. The apparatus of claim 11, further comprising computerized storage, processing and programming for: centering a kernel of said kernel transform.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (22)
Kynor David B. (San Diego CA) Haupt Christopher (San Diego CA) Wilson Steven (Del Mar CA), Artifact removal from physiological signals.
Abraham-Fuchs Klaus (Erlangen DEX) Schlang Martin (Munich DEX) Uebler Johann (Nuernberg DEX), Method for the classification of field patterns generated by electrophysiological activities.
Palazzolo, James Adam; Berger, Ronald D.; Halperin, Henry R.; Sherman, Darren R., Method of determining depth of compressions during cardio-pulmonary resuscitation.
Sharpe Steven M. (Atlanta GA) Seals Joseph (Stone Mountain GA) MacDonald Anita H. (Tucker GA) Crowgey Scott R. (Avondale Estates GA), Non-contact vital signs monitor.
Neuneier, Ralf; Zimmermann, Hans-Georg, System and method for training and using interconnected computation elements to determine a dynamic response on which a dynamic process is based.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.