This Repository was an Implementation code for paper:
https://www.kansaiuniversityreports.com/article/classification-of-normal-and-abnormal-heart-sound-using-continuous-wavelet-transform-and-resnet-50
for further methods and reports, please visit above links.
Heart sound is the sound produced from the mechanical activity of the heart. Some researchers say the sound of the heart occurs from the opening and closing of the heart valve; some researchers say it occurs due to the eddy flow of blood in the heart chamber. Heart in a healthy condition produces certain heart sounds, while an unhealthy heart produces different heart sounds. Various studies have tried to develop a method for classifying heart sounds using digital signal processing methods. The proposed method generally consists of the feature extraction method and classifier. In this study, continuous wavelet transforms and residual neural network (ResNet-50) were used to classify normal and abnormal heart sounds. The lowest error-rate of 0.066 was achieved using 130x130 features. This result was quite competitive compared to previous research. The proposed method is ready to be tested on a dataset with more heart sounds abnormalities
To make this code run without problem, there ware some prerequisites dependency which are:
- Python 3
- Keras and Tensorflow 2
- Librosa
- Pywavelets
- Scipy
- Sklearn
- Achmad Rizal
- Ahmad Alfi Adz-Dzikri - serabiasin
- Muhammad Arik Geraldy Fauzi - ArikGeraldy
This project is licensed under the MIT License - see the LICENSE.md file for details
- Classification of Heart Sound Recordings - The PhysioNet Computing in Cardiology Challenge 2016 - https://physionet.org/content/challenge-2016/1.0.0/