Machine Learning Algorithms for Cardiovascular Disease

Project Summary

In 2021, the United States experienced nearly 700,000 deaths and roughly $250 billion in direct costs due to cardiovascular disease. Current state-of-the-art screening technology, such as electrocardiogram (ECG) and phonocardiogram (PCG) machines, are not available in many medical facilities and are expensive. Auris AI is seeking to develop an AI enhanced digital stethoscope that is easy to use for general practice nurses and doctors and is both portable and inexpensive. Our team set out to develop machine learning (ML) models that can be integrated with Auris AI’s enhanced digital stethoscope and can diagnose cardiovascular diseases using real-time ECG and PCG readings.

Design Goal

Our goal was to conduct research and development of machine learning models which achieve accuracy, while considering the constraints posed by the limited computational resources inherent in an embedded  stethoscope package. Our ML models and research will contribute to the future design of Auris AI’s Enhanced Digital Stethoscope.

Design Constraints

• Model Accuracy
• Model Memory Utilization
• Power Consumption
• Real-time Capability