39 results about "ECG feature" patented technology

Techniques are provided for detecting stroke within a patient using an implantable medical device in conjunction with an external confirmation system. In one example, a preliminary detection of stroke is performed by a subcutaneous monitor based on an analysis of features of an electrocardiogram (ECG) sensed within the patient. Exemplary ECG features indicative of possible stroke include the onset of prominent U-waves, the onset of notched T-waves, and changes in ST segment duration or QT duration or dynamic trends in these parameters. The monitor transmits a signal indicative of possible stroke to a bedside monitor or other external system, which generates a stroke questionnaire for use in confirming the stroke. Family members or other caregivers input answers to the questionnaire into the external system, which confirms or disconfirms the stroke. Emergency personnel can be automatically notified. Implantable systems that detect stroke based on intracardiac electrogram (IEGM) signals are also described herein.

The present invention provides a method and apparatus for determining the occurrence of apnoeas or hypopneas from ECG signal data alone. The method is carried out by apparatus configured to acquire ECG signals from a sleeping subject, transform the signals to data, and extract ECG features relevant to estimate breathing effort for the determination of respiratory events characteristic of apnoeas and hypopneas. The extracted ECG features are correlates of breathing efforts and are used as surrogate measures of breathing or respiratory events. The method may include calculating an AHI or apnoea / hypopnea index. The method may classify apnoeas into obstructive or central apnoeas.

A diagnostic electrocardiogram system employing an electrode lead system (40) for generating one or more electrode signals indicative of electrical activity of a subject heart (10). The diagnostic electrocardiogram system further employs a diagnostic electrocardiograph (50) coupled to the electrode lead system (40) for communicating (e.g., listing, displaying and / or printing) a subject electrocardiogram (20) and one or more diagnostic electrocardiograms (30) designated as a morphology match to the subject electrocardiogram (20). The subject electrocardiogram (20) includes one or more interpretations of ECG features derived from the electrical activity of the subject heart (10) as indicated by the electrode signal(s) (e.g., an algorithmic interpretation and / or an electrocardiographer interpretation of the subject electrocardiogram (20)). The diagnostic electrocardiogram(s) includes one or more diagnoses of ECG features derived from recorded electrical activity of diagnosed heart(s) (11)(e.g., an algorithmic diagnosis and / or an electrocardiographer diagnosis of the diagnostic electrocardiograms(s) (30)).

The present invention discloses at least two electrodes disposed at the body of the ECG measuring device of the present invention. When a user normally uses the ECG measuring device, the two electrodes are just adhered to the hands and the face of the user. The ECG measuring method of the present invention is to measure a first ECG signal from the left hand to the left side of face of the user and a second ECG signal from the right hand to the right side of face of the user. Then a plurality of ECG features are obtained from the first ECG signal, the second ECG signal and the result, as related coefficients, of the interactive computing, as a method of plus and subtract, of the first ECG signal and the second ECG signal. Hence, a biometric authorization process is engaged according to the plurality of ECG features.

The technical solution of the present invention includes a method for extracting electrocardiographic features based on PCANet, which is characterized in that it includes the following steps: S10, performing preprocessing on the electrocardiogram to obtain a training set and a set to be classified, S20, using PCANet to separate the training set and the set to be classified The classification set carries out the feature extraction of the heart beat, S30, train the classifier using the heart beat feature extracted from the training set and use it for the classification of the heart beat feature of the set to be classified; the beneficial effect of the present invention is: it is robust to the noise of the electrocardiogram signal, Simplify the steps of noise removal, have a better classification effect on unbalanced heart beats, improve the efficiency and accuracy of ECG feature extraction, reduce the pressure on doctors to identify ECGs, and reduce the probability of doctors' misdiagnosis.

The invention discloses a pre-diabetes detection system and method based on the combination of electrocardiogram and electroencephalogram information. The system includes: a signal acquisition module, which uses a wearable device to obtain the user's ECG signal and EEG signal synchronously in a non-invasive manner; Dimensionality reduction processing is performed on multiple dimensionality reduction sets to obtain multiple dimensionality reduction combination feature sets, and by analyzing the correlation between multiple dimensionality reduction combination feature sets and blood glucose concentration values, ECG features and EEG features that meet the set correlation criteria are selected. , to form an optimized combined feature set; the multi-modal fusion module is used to input the optimized combined feature set to the trained neural network models of various types, and obtain the user's diabetes by fusing the output results of various types of neural networks. Preliminary test results. The invention can provide a non-invasive, painless, convenient, comfortable and low-cost pre-diabetes detection solution.

The electrocardiogram feature selection method provided by the present invention divides the electrocardiogram into four types of NSVF classification systems, and decomposes the four types of NSVF classification systems into six binary classifiers of NvS, NvV, NvF, SvV, SvF, and VvF. In the classifier, each feature is sorted according to the score to form a candidate feature set, and then the optimal feature subset is selected from each binary classifier, and the ECG sample to be tested is predicted based on the optimal feature subset to obtain the ECG to be tested kind of category. The electrocardiogram feature selection method provided by the present invention forms a feature subset after sorting the feature scores from high to low, and selects the optimal feature subset from each binary classifier, and uses the optimal feature subset to predict the ECG sample to be tested, The category of the electrocardiogram to be tested is obtained, and the prediction accuracy is improved.

The invention discloses a fatigue detection method based on electroencephalogram and electrocardiogram in which the steering wheel is embedded with an electrocardiogram sensor. Fatigue detection is performed by constructing a product fuzzy convolution network, which specifically includes: S1. Obtaining electrocardiogram data through an electrocardiogram detection chip, and using The electroencephalograph obtains EEG time series data; S2, uses a fuzzy neural network with multilayer feedback to process EEG time series data, and obtains EEG features; S3, builds a deep network based on a one-dimensional convolutional neural network framework The feature extraction network extracts the fatigue features of the ECG data and generates the ECG feature sequence; S4. Design the fusion network, input the ECG feature sequence and the EEG feature at the same time, fuse the two signals together, and give the predicted value; S5 , Use the adaptive moment estimation algorithm to optimize and train the network model. It can reduce noise and improve detection accuracy. The introduction of multilayer reduces the limitation of fuzzy neural network on the feature dimension of input data and improves the accuracy of classification results.

The invention provides an ECG (electrocardiograph) feature extraction method integrating Butterworth filtering and wavelet transformation, and belongs to the technical field of medical computers. A core concept is described as follows: performing de-noising processing on power frequency interference of an original ECG signal; meanwhile, removing baseline drift in a curve fitting manner; respectively positioning a wave Q, a wave R and a wave S with a differential threshold method and a wavelet transformation method after a relatively accurate ECG signal is acquired, comparing the waves Q, R andS with information marked by an expert, and respectively regulating threshold, scale and translation amount parameters to obtain an optimal positioning result. Based on the result, the wave Q, the wave R and the wave S are positioned respectively by adopting the differential threshold method or the wavelet transformation method. According to the extraction method provided by the invention, the more accurate de-noising processing can be performed on the original ECG signal; in addition, the waves Q, R and S can be positioned accurately, therefore, a foundation is laid for a digital analysis onan ECG.