Electrocardiogram signal segmentation
AssigneeMedicalgorithmics Ltd.
InventorMarek Dziubinski
Life of this patent
Estimating the term…
Abstract
Techniques are disclosed for segmenting electrocardiogram (ECG) signals. In one example, a method to segment an electrocardiogram (ECG) signal may include detecting consecutive heartbeats in an ECG signal. The method also includes segmenting the ECG signal into multiple ECG segments surrounding the detected consecutive heartbeats and generating an ECG data set by joining consecutive ECG segments. The generated the ECG data set represents the detected heartbeats. In some such examples, each ECG segment is of a duration to include a QRS complex, a P wave, and a T wave.
Claims19
A method to segment an electrocardiogram (ECG) signal, the method comprising:
- detecting, by one or more processors, a plurality of consecutive heartbeats in an ECG signal;
- segmenting, by the one or more processors, the ECG signal into a plurality of ECG segments surrounding the detected plurality of consecutive heartbeats;
- generating, by the one or more processors, an ECG data set by joining consecutive ECG segments, wherein the ECG data set represents the detected plurality of consecutive heartbeats;
- detecting, by the one or more processors, one or more adjacent pairs of ECG segments within the ECG data set that do not overlap in time;
- for each of the pairs of non-overlapping ECG segments, including an artificial heartbeat within the ECG data set to indicate a missing ECG fragment between the pair of non-overlapping ECG segments, wherein the artificial heartbeat is a marker between the pair of non-overlapping ECG segments;
- generating, by the one or more processors, a machine learning input data table using the ECG data set representing the detected plurality of consecutive heartbeats and including the artificial heartbeats, such that the machine learning data table includes a continuous representation of ECG data;
- detecting one or more emergent events or information regarding detected non-emergent events based on the machine learning input data table;
- and
- generating a report based on the detections, wherein the report includes information regarding the detected one or more emergent events or information regarding detected non-emergent events.
A system to segment an electrocardiogram (ECG) signal, the system comprising:
- one or more non-transitory machine-readable mediums configured to store instructions;
- and
- one or more processors configured to execute the instructions stored on the one or more non-transitory machine-readable mediums, wherein execution of the instructions causes the one or more processors to:
- detect, by the one or more processors, N consecutive heartbeats in an ECG signal;
- segment, by the one or more processors, the ECG signal into a plurality of ECG segments surrounding the detected N consecutive heartbeats;
- generate, by the one or more processors, an ECG data set by joining consecutive ECG segments, wherein the ECG data set represents the detected N consecutive heartbeats;
- detect, by the one or more processors, one or more adjacent pairs of ECG segments within the ECG data set that do not overlap in time;
- for each of the pairs of non-overlapping ECG segments, including an artificial heartbeat within the ECG data set to indicate a missing ECG fragment between the pair of non-overlapping ECG segments, wherein the artificial heartbeat is a marker between the pair of non-overlapping ECG segments;
- generate, by the one or more processors, a machine learning input data table using the ECG data set representing the detected N consecutive heartbeats and including the artificial heartbeats, such that the machine learning input data table includes a continuous representation of ECG data;
- detect one or more emergent events or information regarding detected non-emergent events based on the machine learning input data table;
- and
- generate a report based on the detections, wherein the report includes information regarding the detected one or more emergent events or information regarding detected non-emergent events.
Dependencies are inferred from the claim text.