Huang, Yi, and Song, Insu (2018) A better online method of heart diagnosis. In: Proceedings of the 3rd International Conference on Biomedical Signal and Image Processing. pp. 81-86. From: ICBIP 2018: 3rd International Conference on Biomedical Signal and Image Processing, 22-24 August 2018, Seoul, Korea.

PDF (Published Version) - Published Version Restricted to Repository staff only

Abstract

Heart related health conditions, such as heart attacks, are the primary reasons for millions of deaths worldwide. Most of the existing heart-disease diagnosis techniques are invasive and need trained medical professionals. The current paper reports the development of a rapid, non-invasive heart sound diagnosis method and describes its functionality in Health Social Networks (HSNs). HSNs are platforms for health professionals, organizations and patients to use IT applications to share information, including medical information, e.g., HSNs allow users to conveniently and inexpensively upload their heart sounds for professional analysis using mobile phones. This method offers great promise, as heart-related health conditions cause millions of deaths worldwide. This paper contributes to the development of a new feature extraction method based on Gaussian Hamming Distance (GHD). The approach that applying GHD to MFCC features (GMTS) is shown robust to noise and also solves the difficulties of recording heart sound activities. The proposed GMTS feature was tested with 0.05 seconds shifting stress and obtained accuracy, sensitivity and specificity of at least 84.2%, 77.8%, and 84.5%, respectively.

Actions (Repository Staff Only)

Item Control Page