Heart sound classification using deep structured features

Authors

Michael Tschannen, Thomas Kramer, Gian Marti, Matthias Heinzmann, and Thomas Wiatowski

Reference

Computing in Cardiology (CinC), Vancouver, Canada, pp. 565-568, Sept. 2016.

Abstract

We present a novel machine learning-based method for heart sound classification which we submitted to the PhysioNet/CinC Challenge 2016. Our method relies on a robust feature representation--generated by a wavelet-based deep convolutional neural network (CNN)--of each cardiac cycle in the test recording, and support vector machine classification. In addition to the CNN-based features, our method incorporates physiological and spectral features to summarize the characteristics of the entire test recording. The proposed method obtained a score, sensitivity, and specificity of 0.812, 0.848, and 0.776, respectively, on the hidden challenge testing set.

Comments

Code for the scattering transform is available here.

Download this document:

This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.