Arrhythmia identification using deep learning to improve accuracy of detection and faster treatment of condition

Abnormality detection in heart rates
Customer Challenge

Arrhythmia, or irregular heartbeat, is a problem when the rate or rhythm with which the heart beats, is not normal. The healthcare startup is building a product that can detect and predict heart conditions based on the 15 different arrythmia patterns based on patients ECG (electrocardiography) and PPG (photoplethysmography) data.

Solution Implemented

Building and training Neural Network models that can learn from and detect hearth conditions based on arrhythmic ECG and PPG data and testing the new models against the datasets as specified in the ANSI/AAMI EC57: 2012/®2020 document. The solution will help the customer with FDA approval process to commercialize the solution.

Outcome Delivered

This is an ongoing project, and the success of the project will be measured from two perspective:

  • Approval of the models and the solution by FDA
  • The models are OEM-ed into heartrate monitoring and detection systems