Heartbeat Sound Segmentation and Classification

• According to WHO, Cardiovascular Diseases (CVD's) continue to be one of the leading causes of deaths globally.
• To check for any CVD's (abnormalities) in patients' heartbeat sounds, medical practitioners currently use a method known as cardiac auscultation.
• This is a process whereby a medical practitioner listens to the heart sound, analyses it and classifies it as normal or abnormal.
• Generally it is a difficult skill to acquire considering the complexity of abnormal heart sounds.
• An easily accessible and reliable heartbeat sound classification system would be vital in reducing high mortality rates due to CVD's and also assist medical practitioners with more accurate cardiac auscultation.

• To implement a method which can locate lub and dub sounds (S1 and S2) within audio data, segment the files and classify heartbeats into normal or diseased categories.
• To create a model that will enable a first level screening of detecting abnormalities in an individuals heart sound.

  For home use by individuals using a smartphone.
  For hospital use by medical professionals.

The below diagram presents the project methodology.