Monitoring of Sleep Breathing States Based on Audio Sensor Utilizing Mel-Scale Features in Home Healthcare

Area of Science:

• Biomedical Engineering
• Signal Processing
• Sleep Medicine

Background:

• Sleep-related breathing disorders (SBDs) significantly impair sleep quality and elevate risks of severe cardiovascular and cerebrovascular diseases.
• Accurate detection of SBDs is crucial for timely intervention and preventing life-threatening complications.

Purpose of the Study:

• To develop and validate a novel method for detecting SBDs using breathing sound signals.
• To differentiate between normal and abnormal sleep breathing states for improved SBD diagnosis.

Main Methods:

• Moment waveform analysis was employed for precise localization and segmentation of breathing cycles.
• Mel frequency cepstrum analysis was utilized to extract key features from breathing signals.
• Extracted Mel-scale indexes were used to classify breathing states as normal or abnormal.

Main Results:

• The proposed method demonstrated high efficacy in distinguishing normal and abnormal sleep breathing states.
• An average accuracy of 93.1% was achieved in detecting abnormal breathing states.
• The method was successfully tested on both healthy individuals and patients with obstructive sleep apnea.

Conclusions:

• The developed breathing sound analysis technique offers a reliable approach for SBD detection.
• This method holds significant potential for home healthcare applications, facilitating early SBD prevention and sleep quality enhancement.