An Algorithm to Analyze the Heart Sound to Diagnose Heart Diseases
摘要
This paper proposes a new method for analyzing heart sounds, focusing on detecting and measuring the temporal intervals between the two primary heart sounds: the first heart sound (S1, Lubb) and the second heart sound (S2, Dub). The primary objectives of this work are to accurately count heartbeats, quantify the time gaps between these significant cardiac events, and calculate the heart rate. The proposed method employs a peak detection-based approach to reliably identify S1 and S2 sounds. Two experiments were conducted and evaluated. The first achieved an accuracy of 95.6% when applied to high-quality heart sounds collected from 15 diverse donors of varying ages and genders. The second experiment, carried out on a publicly available dataset with low-quality heart sounds, recorded a drop in accuracy to 75.194%. Moreover, bar graphs were used to clearly and efficiently represent the results of the analysis.