📄 中文摘要
打鼾常被视为夜间的笑料，但对于数百万人来说，它是阻塞性睡眠呼吸暂停（OSA）的症状，这是一种严重的疾病，导致呼吸反复停止和恢复。开发者可以利用智能手机将其转变为诊断级监测工具。该教程深入探讨音频信号处理和深度学习，构建OSA筛查工具。通过使用OpenAI Whisper进行强大的音频去噪，Librosa进行特征提取，以及经过微调的PyTorch卷积神经网络（CNN）来分类呼吸模式，旨在为对医疗AI、音频深度学习或边缘计算感兴趣的开发者提供指导。
超越打鼾：使用 OpenAI Whisper 和 PyTorch 进行实时睡眠呼吸暂停筛查
📄 English Summary
Beyond the Snore: Real-time Sleep Apnea Screening with OpenAI Whisper and PyTorch
Snoring is often dismissed as a late-night joke, yet for millions, it signifies Obstructive Sleep Apnea (OSA), a serious condition characterized by repeated interruptions in breathing. Developers have the capability to transform smartphones into diagnostic-grade monitors. This tutorial delves into Audio Signal Processing and Deep Learning to create an OSA screening tool. It utilizes OpenAI Whisper for effective audio denoising, Librosa for feature extraction, and a fine-tuned PyTorch CNN to classify breathing patterns. This guide aims to assist those interested in AI in Healthcare, Deep Learning for Audio, and Edge Computing.