Urvi Patel with Dr. Ronald Cechner (Anesthesiology, University Hospitals)

Sound Analysis to Automatically Distinguish Wheezes, Stridor, and Crackles

Although the presence of wheezes, crackles, and stridor sounds indicate obstructed airways and other lung problems, there currently exists no method to continuously monitor breath sounds for these abnormalities during surgery. The goal of this project is to build a non-obstructive device that can continuously monitor breath sounds in order to detect and accurately identify wheezes, crackles, and stridor sounds. We seek to record high quality breath sounds using a microphone attached to an LMA, endotracheal tube, or mask in order to automatically analyze these sounds in real time. In addition, various techniques in signal analysis, including Fourier and wavelet transforms, will be explored to efficiently distinguish the three key breath sounds (wheezes, stridor, and crackles). Finally, if possible, we hope to test the device in a clinical setting.