Phonocardiogram Signal Analysis for Automated Cardiac Murmur Detection

Valvular diseases cause a crucial health burden today, necessitating effective diagnostic tools and modern methods. Automated techniques play a vital role in the early diagnosis and detection of heart diseases with the critical approach in modern healthcare. These automated methods employ deep learning approaches and signal processing to analyze cardiac data efficiently and accurately to preprocess biomedical signals, such as phonocardiograms. Preprocessing techniques like windowing, variation calculation, and denoising are pivotal in enhancing signal quality and extracting relevant features for disease classification. This paper presents a PCG signal preprocessing and spectrogram generation framework, followed by murmur classification as present or absent. Our results highlight the importance of preprocessing techniques in improving the reliability and efficacy of automated heart disease detection systems.