Heart Sounds Classification Using Loudness Features and Gaussian Mixture Model
The heart sound’s loudness characteristics are used in this paper to introduce a new automatic way of identifying the status of the heart sound. The three main steps of the approach are as follows. First, an 8th order Chebyshev Type-II filter is used to heavily filter the heart sound, which is typically found to be loud. The filtered heart sound is then divided into the first heart sound, systole, second heart sound, and diastole using the event synchronous approach. The loudness attribute is expressed using the mean rows of its spectrogram in the second stage. The Gaussian Mixture Model method is used in the third phase to categorize the heart sound. The suggested method has been tested on a large database of heart sounds with over 3000 recordings, and it has a success rate of 97.77 percent.
Vishwanath Madhava Shervegar,
E&CE Department, Mangalore Institute of Technology & Engineering, Moodbidre, DK, Karnataka, India.
Please see the link here: https://stm.bookpi.org/CPMS-V5/article/view/7669
Keywords: Phonocardiography, event synchronous segmentation, loudness, spectrogram, GMM classifier