Abnormal Heart Beat Detection Based on Deep Learning

Description Machine learning techniques can reduce burden using manual auscultation for diagnosing heart disease.  In this regard, this project investigates the efficacy of using deep learning for the classification of Phonocardiogram (PCG) signals.
Task Using scalogram images of PCG signals the students first will develop Convolutional Neural Networks (CNNs) structures for the task of heart sound classification.
Utilises Python and related deep learning toolkits (pytorch/keras/tflearn/caffe)
Requirements Basic machine learning knowledge, python knowledge
Languages English
Supervisor Zhao Ren, M. Sc. (