Listening to heartbeats was the original method of diagnosing cardiac anomalies. While some anomalies can be detected by analysing ECG data, some cardiac flow anomalies are still initially diagnosed by listening to heartbeats.
This project aims to identify such anomalies by first isolating heartbeat sounds from audio input, then classifying these sounds as normal or one of several occurring anomalies.
To see code and results, follow the link to my Github page.