Deep learning-based snore sound analysis for the detection of night-time breathing disorders
Bo Dang, Danqing Ma, Shaojie Li, Zongqing Qi, Elly Yijun Zhu
Abstract
Snoring, a prevalent symptom of obstructive sleep apnea, is believed to impact 57% of men and 40% of women in the United States. Night-time breathing disorders present significant challenges to both diagnosis and treatment, impacting millions of individuals worldwide. Traditional methods like CPAP machines and lifestyle changes face barriers such as discomfort, low adherence, and high costs, prompting the need for innovative solutions. This paper presents a novel approach using artificial intelligence, specifically deep learning, to create a snore sound analysis-based alerting system. This system aims to detect sleep disorders by analyzing snore patterns, providing a non-intrusive, cost-effective, and user-friendly alternative to traditional methods. By training models on snore sound characteristics, we've achieved promising results in identifying sleep apnea, showcasing the potential of this system in transforming the detection and management of night-time breathing disorders.