top of page

Problem Statement:  Several components of heart sounds carrying clinical  importance fall beyond human audible range. Make algorithms to detect these faint sounds from a stethoscope signal for primary care clinical decision support.​

​Solution highlights:

​

  • Apart from the normal S1|S2, also identify S3, S4 sounds, Split S1, Systolic and Diastolic Murmurs

  • Provide clinical decision support powered by AI

  • Triage using stethoscope signal recorded from multiple positions on chest

​Tech Stack used:

​

Acoustic Signal Analysis, Wavelet Transforms, Time-Delay Neural Networks, Mel-Frequency Cepstral Coeffs, Hidden Markov Models.

Previous 

Hand washing Training and Compliance

wix datakriti (10)_edited.jpg

Next

Energy Disaggregation

bottom of page