Convolutional Long Short-term memory ( CNN - LSTM ) evaluation on the heart sound database with 91% accuracy.
My submission on the 2016 abnormal heart sound classification challenge.
More info can be found on https://physionet.org/content/challenge-2016/1.0.0/.
Convolutional Long Short-term memory evaluation on the heart sound database with 91% accuracy.
Classification was done on raw data. Only a band pass butter worth was applied for preprocessing.