Heartbeat Sound Classification Using Deep Learning Techniques Over Raspberry Pi System

Abstract

This article focuses on the development of a heart beat sound classification system utilizing sound signals through the combined application of Convolutional Neural Networks (CNN) techniques over Raspberry Pi System. The proposed methodology involves the acquisition of heart sounds, preprocessing through signal analysis using CNN model accurate classification. By utilizing the stated algorithms, the aim is to determine task of heart rate classification.The core system leverages CNN model to effectively classify heart beat sounds. A diverse dataset comprising a range of heart conditions and rates is used to train and fine-tune both deep learning architectures. This approach allows for the extraction of intricate features from the audio data, enabling improved classification accuracy.The proposed system classifies heart beat sounds into categories such as normal, murmur, extra heart sound, extrasystole, and artifact. Through extensive training and validation, CNN model will learn to recognize distinctive patterns in the audio signals, facilitating precise classification of different heart beat sounds. Upon evaluation, the performance of CNN algorithm will be compared to determine the results in heart sound classification. The whole analysis process is carried on a Raspberry Pi machine,  which is very cost-effective and portable device. The Raspberry Pi serves as an accessible platform for real-time heartbeat sound classification, making advanced cardiac monitoring technology more widely available. This comparative analysis will contribute valuable insights into the effectiveness of each approach in this specific application.