Current technologies analyze electrocardiogram (ECG) signals for a long duration, which is not always a practical scenario. Moreover the current scenarios perform a binary classification between normal and Atrial Fibrillation (AF) only, whereas there are many abnormal rhythms apart from AF. Conventional systems/methods have their own limitations and may tend to misclassify ECG signals, thereby resulting in an unbalanced multi-label classification problem. Embodiments of the present disclosure provide systems and methods that are robust and more efficient for classifying rhythms for example, normal, AF, other abnormal rhythms and noisy ECG recordings by implementing a spectrogram based noise removal that obtains clean ECG signal from an acquired single-lead ECG signal, an optimum feature selection at each layer of classification that selects optimum features from a pool of extracted features, and a multi-layer cascaded binary classifier that identifies rhythms in the clean ECG signal at each layer of the classifier.

An artificial intelligence self-learning-based static electrocardiography analysis method and apparatus, the method comprising data preprocessing, heartbeat detection, heartbeat classification based on a depth learning method, heartbeat verification, heartbeat waveform feature detection, measurement and analysis of electrocardiography events, and finally automatic output of reporting data, realizing an automated static electrocardiograph analysis method having a complete and rapid flow. The static electrocardiography analysis method can also record modification information of an automatic analysis result, collect modified data, and feed same back to the depth learning model to continue training, thereby continuously making improvements and improving the accuracy of the automatic analysis method.

A self-learning dynamic electrocardiography analysis method employing artificial intelligence. The method comprises: pre-processing data, performing cardiac activity feature detection, interference signal detection and cardiac activity classification on the basis of a deep learning method, performing signal quality evaluation and lead combination, examining cardiac activity, performing analytic computations on an electrocardiogram event and parameters, and then automatically outputting report data. The method achieves an automatic analysis method for a quick and comprehensive dynamic electrocardiography process, and recording of modification information of an automatic analysis result, while also collecting and feeding back modification data to a deep learning model for continuous training, thereby continuously improving and enhancing the accuracy of the automatic analysis method. Also disclosed is a self-learning dynamic electrocardiography analysis device employing artificial intelligence.

Provided are an electrocardiogram signal parallel analysis apparatus, a mobile terminal incorporating the apparatus, and related methods. The apparatus includes an integrated memory, a central processing unit and a graphic processing unit. The integrated memory includes a first memory and a second memory for being used by the central processing unit and the graphic processing unit respectively, and the central processing unit may access the second memory. The central processing unit performs primary noise reduction on a received electrocardiogram original signal to obtain a primary electrocardiogram signal, and performs abnormal heartbeat classification preliminary screening on characteristic data extracted from the graphic processing unit to obtain suspected abnormal heartbeat data. The graphic processing unit performs characteristic extraction on the primary electrocardiogram signal to obtain characteristic data, performs secondary noise reduction on the primary electrocardiogram signal to obtain a secondary electrocardiogram signal, and processes the suspected abnormal heartbeat data and the secondary electrocardiogram signal by applying a template matching classification mode to obtain final abnormal heartbeat data.

An ambulatory medical device including a plurality of sensing electrodes and one or more processors operably coupled to the plurality of sensing electrodes is provided. Each sensing electrodes is configured to be coupled eternally to a patient and to detect one or more ECG signals. The one or more processors are configured to receive at least one electrode-specific digital signal for each of the plurality of sensing electrodes, determine a noise component for each of the electrode-specific digital signals, analyze each of the noise components for each of the plurality of sensing electrodes, generate electrode matching information for each sensing electrode of the plurality of sensing electrodes based upon analysis of each of the noise components, determine one or more sensing electrode pairs based upon the electrode matching information, and monitor each of the one or more sensing electrode pairs for ECG activity of the patient.

Described herein is a system and method of automatically monitoring QT intervals in a patient based on one or more EKG signals received from attached monitoring devices. Each EKG signal is analyzed to detect attributes of the first and second EKG signals, including QRS onset information, QRS peak information, and T-wave offset information. A QT interval is calculated based on QRS onset information derived from the first EKG signal and T-wave offset information derived from the second EKG signal. The calculated QT interval is compared to thresholds to detect elongation of the QT interval and an alert is generated in response to a detected elongated QT interval.

There is provided an ambulatory system, comprising at least first and second wearable sensors, for determining pulse transit time (PTT) between at least a first and at least a second fixed location within the cardiovascular system of a subject. The system comprises at least a first device, wherein the first device can contact the skin of the subject, the first device being positioned proximate to the first fixed location; and also comprises at least a second device, wherein the second device can contact the skin of the subject, the second device being positioned proximate to the second fixed location. The system further comprises a data collection module that is in communication with the first and second devices. The first device is configured to detect a timing cue within the cardiac cycle of the subject, and the second device is configured to detect a pulse pressure wave passing through the second fixed location. The data collection module collects data relating to the transition of the pulse pressure wave passing through the second fixed location, thereby enabling determination of a pulse transit time (PTT) between the first and second fixed locations.

A method for vector analysis of an electrocardiogram for assessment of risk of sudden cardiac death includes receiving data about electrical activity of heart of a subject recorded on electrocardiogram device, generating a vector cardiogram based on the data, analyzing the vector cardiogram to determine arrhythmogenic right ventricular dys-plasia/cardiomyopathy to identify a presence of a micro-scar in a three-dimensional vector loop of the vector cardiogram, determining a risk of SCD for the subject based on the identification of the presence of a micro-scar, and storing the risk in a database.

A system is provided for isolating the value of a signal hidden within a composite electrical signal. The system comprises an input, a processor, and a memory configured to store instructions executable by the processor. The instructions cause the processor to estimate that portion of a received composite electrical signal that represents a hidden signal by subtracting a known first signal from the composite signal.

Some embodiments are directed to a device for processing a fetal electrocardiogram (fECG). The processing may include obtaining a fetal vector cardiogram (VCG) from said multiple ECG signals, and providing the fetal vector cardiogram as an input to the machine learning classifier configured with trained parameters, and obtain a classification from the machine learning classifier.