Development of an Electronic Stethoscope and a Classification Algorithm for Cardiopulmonary Sounds

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Medicine
• Medical Diagnostics

Background:

• Conventional stethoscopes have limitations in auscultation accuracy due to subjective interpretation and inability to record sounds.
• Variability in physician training and age-related hearing decline can impact diagnostic consistency.
• Objective analysis of cardiopulmonary sounds is crucial for accurate diagnosis and treatment.

Purpose of the Study:

• To develop an electronic stethoscope integrated with an AI-based classifier for objective cardiopulmonary sound analysis.
• To overcome the limitations of traditional stethoscopes in sound recording and interpretation.
• To improve the accuracy and consistency of diagnosing heart and lung conditions.

Main Methods:

• Development of an electronic stethoscope with an embedded condenser microphone and optimized noise reduction circuits.
• Application of Fast Fourier Transform (FFT) for analyzing microphone placement and noise reduction strategies.
• Implementation of AI for classifying cardiopulmonary sounds using Mel-frequency cepstral coefficients (MFCC) and ensemble learning on segmented sound frames.

Main Results:

• The microphone placement surrounded by cork demonstrated superior noise reduction.
• Distinct AI classifiers were developed for heart and lung sounds, achieving high performance metrics.
• Optimal performance for heart sound classification included 86.9% accuracy, 81.9% sensitivity, and 91.8% specificity.
• Lung sound classification achieved 73.3% accuracy, 66.7% sensitivity, and 80% specificity.

Conclusions:

• The developed electronic stethoscope with AI classification offers a significant advancement over conventional methods.
• The system provides objective and reliable analysis of cardiopulmonary sounds, aiding in clinical decision-making.
• This technology has the potential to enhance diagnostic accuracy for various heart and lung pathologies.