Method and system for cardiac signal processing
A system for cardiac signal processing, preferably including any or all of: a data collection device, a set of computing and/or processing subsystems, a set of algorithms and/or models, and/or a set of output devices. A method for cardiac signal processing, preferably including processing a set of inputs to determine a set of metrics and determining a set of outputs based on the set of metrics and/or a set of supplementary metrics, and optionally including any or all of: receiving a set of inputs, determining a set of supplementary metrics associated with the set of metrics, and/or triggering the set of outputs.
1. A method for cardiac signal processing, comprising:
receiving cardiac data associated with a patient;
determining that the cardiac data is valid for use with a trained detection model, comprising providing the cardiac data to a trained quality model and a set of criteria-based models, wherein the trained quality model is trained to detect low quality input data within the cardiac data based on a set of quality indices, wherein the low quality input data is not valid for use with the trained detection model, wherein the set of criteria-based models validates the cardiac data based on a signal format of the cardiac data and a set of signal characteristics of the cardiac data, wherein the trained detection model is associated with detection of a medical condition;
processing the cardiac data with the trained detection model to produce an output indicative of the medical condition.
2. The method of claim 1 , wherein the cardiac data comprises an electrocardiogram (ECG) signal, wherein the set of signal characteristics of the cardiac data comprise at least one of: ECG signal duration, ECG signal sampling rate, or ECG lead corruption, wherein the signal format of the cardiac data comprises a tensor format.
3. The method of claim 1 , wherein determining that the cardiac data is valid for use with the trained detection model is performed with a trained quality model.
4. The method of claim 3 , wherein the trained quality model is separate and distinct from the trained detection model.
5. The method of claim 1 , wherein the trained detection model comprises a set of neural networks.
6. The method of claim 5 , wherein the trained detection model comprises an ensemble of trained neural networks.
7. The method of claim 6 , wherein the trained detection model further comprises an aggregator.
8. The method of claim 1 , further comprising providing the output to a set of users.
9. The method of claim 1 , wherein the output is determined based on a prediction metric produced by the trained detection model, wherein the prediction metric is indicative of a likelihood that the patient has the medical condition.
10. The method of claim 9 , wherein producing the output further comprises comparing the prediction metric with a threshold value.
11. A system for cardiac signal processing, comprising:
a trained detection model associated with detection of a medical condition;
a trained quality model trained to detect low quality input data within cardiac data based on a set of quality indices, wherein the low quality input data is not valid for use with the trained detection model;
a set of criteria-based models configured to validate the cardiac data based on a signal format of the cardiac data and a set of signal characteristics of the cardiac data;
a processing subsystem configured to:
receive the cardiac data associated with a patient;
determine, with the trained quality model and the set of criteria-based models, that the cardiac data is valid for use with the trained detection model; and
process the cardiac data with the trained detection model to produce an output indicative of the medical condition.
12. The system of claim 11 , wherein the cardiac data comprises an electrocardiogram (ECG) signal, wherein the set of signal characteristics of the cardiac data comprise at least one of: ECG signal duration, ECG signal sampling rate, or ECG lead corruption, wherein the signal format of the cardiac data comprises a tensor format.
13. The system of claim 11 , wherein the trained detection model comprises a set of neural networks.
14. The system of claim 13 , wherein the trained detection model comprises an ensemble of trained neural networks.
15. The system of claim 14 , wherein the trained detection model further comprises an aggregator.
16. The system of claim 11 , wherein the processing subsystem is further configured to provide the output to a set of users.
17. The system of claim 16 , wherein the set of users comprises multiple users.
18. The system of claim 11 , wherein the medical condition is hypertrophic cardiomyopathy.
19. The system of claim 11 , wherein the output is produced based on a prediction metric produced by the trained detection model, wherein the prediction metric is indicative of a likelihood that the patient has the medical condition.
20. The system of claim 19 , the output is further produced based on comparing the prediction metric with a threshold value.