[Recognition and classification of heart sounds]

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Heart sound analysis is crucial for non-invasive cardiac diagnosis.
• Accurate recognition and classification of heart sounds remain challenging.
• Advancements in computational methods offer new possibilities for heart sound analysis.

Purpose of the Study:

• To provide a comprehensive review of current methods for heart sound recognition and classification.
• To highlight recent research advances in the field.
• To discuss the primary challenges and limitations of existing techniques.

Main Methods:

• Review of statistical techniques for heart sound analysis.
• Examination of artificial neural network applications in heart sound classification.
• Analysis of signal processing and feature extraction methods.

Main Results:

• Statistical methods offer robust feature extraction for heart sounds.
• Artificial neural networks demonstrate high accuracy in classifying various heart sound patterns.
• Significant progress has been made in automated heart sound analysis.

Conclusions:

• Both statistical and artificial neural network approaches are vital for heart sound recognition.
• Addressing current methodological problems is key to improving diagnostic accuracy.
• Further research is needed to refine these techniques for clinical application.