A snore extraction method from mixed sound for a mobile snore recorder.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Sleep Medicine

Background:

• Snoring is common in shared sleep environments, often masking individual sound patterns.
• Distinguishing individual snores is crucial for accurate sleep disorder assessment.
• Existing methods struggle with complex acoustic mixtures in home settings.

Purpose of the Study:

• To develop a novel snore recorder capable of separating individual snores from delayed acoustic mixtures.
• To enable detailed analysis of snoring in naturalistic, multi-person sleep conditions.
• To provide a tool for early detection and monitoring of sleep-related breathing disorders.

Main Methods:

• Utilized blind source separation techniques to address the delayed source separation problem.
• Developed a specialized snore separator algorithm.
• Introduced a performance index to track algorithm convergence.

Main Results:

• The developed snore separator effectively isolates individual snores from mixed sound recordings.
• Experimental validation demonstrated the system's good performance in separating complex snore signals.
• The system successfully handles delayed mixtures common in multi-occupant sleep environments.

Conclusions:

• The snore separator offers a viable solution for analyzing snoring in real-world, multi-person sleep settings.
• This technology can facilitate pre-polysomnography screening for sleep apnea symptoms.
• It serves as a valuable post-polysomnography monitoring device for tracking snoring patterns.