Deep learning techniques for digital cardiac auscultation analysis - Hands on session
Date:
I was responsible for the design of the hands-on part of the session led by Prof. Francesco Renna in an EUGLOH BIP organized by the University of Porto on Advanced Processing of Biomedical Signals and Images. The original title was “Deep learning techniques for digital cardiac auscultation analysis”.
This tutorial was designed to equip the advanced practioneer with the fundamental signal processing and deep learning concepts in order to understand and succesfully tackle tasks related to phonocardiogram heart sound segmentation and characterization.
Download part 1 - Data and Signal Processing
Download part 2 - Model serving, heart sound segmentation and murmur classification
The tutorial was designed to be ran in Google Collab (requires at least 16 GB RAM) but can be easily adapted to run locally.