Classification of severity of mitral regurgitation patients using multifractal analysis

Owing to its efficiency and cost effectiveness, auscultation using a stethoscope is the primary diagnosis method preferred by most medical examiners for cardiovascular disease. Phonocardiography (PCG) is the next level of auscultation and an effective analysis technique in cardiology. This paper presents a method to quantify Mitral Regurgitation (MR) using PCG. MR is one of the most common cardiovascular diseases associated with a defect in mitral valve. We analyzed the multifractal nature of heart sounds with murmur followed by classification and clustering using Hurst exponents. Thus, we were able to project the difference between severities of MR patients by correlating the complexity of heart sounds from four major auscultation sites, using a complexity analysis tool called singularity spectra. The simulation results show that the method can quantify the severity of MR using PCG.