Deep neural network on ECG Poincare plot for atrial fibrillation classification

A computer system for detecting and classifying atrial fibrillation (AFib), the computer system comprising:

• one or more computer-readable non-transitory storage media configured to store computer program code;
• and
• one or more computer processors configured to access said computer program code and operate as instructed by said computer program code, said computer program code including:
• receiving code configured to cause the one or more computer processors to receive data corresponding to an electrocardiogram (ECG) associated with a patient;
• extraction code configured to cause the one or more computer processors to extract one or more RR intervals from the data;
• determining code configured to cause the one or more computer processors to determine one or more previous RR intervals for each of the extracted RR intervals;
• aggregating code configured to cause the one or more computer processors to aggregate one or more features associated with the one or more RR intervals and the one or more previous RR intervals into a two dimensional array comprising a first dimension, representing a number of the one or more RR intervals and the one or more previous RR intervals, and a second dimension representing a size of a lookback window of at least one of the one or more RR intervals;
• classifying code configured to cause the one or more computer processors to classify one or more patterns associated with the aggregated features and based on a neural network that is trained based on ECG training data;
• pattern determining code configured to cause the one or more computer processors to determine whether the data contains a pattern associated with AFib based on classifying the one or more patterns;
• and
• controlling code configured to cause the one or more computer processors to control a warning to be output based on determining that the data contains the patter associated with AFib based on classifying the one or more patterns.

A non-transitory computer readable media having stored thereon a computer program for detecting and classifying atrial fibrillation (AFib), the computer program configured to cause one or more computer processors to:

• receive data corresponding to an electrocardiogram (ECG) associated with a patient;
• extract one or more RR intervals from the data;
• determine one or more previous RR intervals for each of the extracted RR intervals;
• aggregate one or more features associated with the one or more RR intervals and the one or more previous RR intervals into a two dimensional array comprising a first dimension, representing a number of the one or more RR intervals and the one or more previous RR intervals, and a second dimension representing a size of a lookback window of at least one of the one or more RR intervals;
• classify one or more patterns associated with the aggregated features and based on neural network that is trained based on ECG training data;
• determine whether the data contains a pattern associated with AFib based on classifying the one or more patterns;
• and
• controlling a warning to be output based on determining that the data contains the patter associated with AFib based on classifying the one or more patterns.