Monitoring apnea syndrome millimeter wave system cracks the "digital code" of snoring
Medical profession
Science and Technology Daily News (Huang Lingyi, Cheng Zhenwei, Qian Sheng, reporter Jiang Yun) Recently, the reporter learned that Hangzhou Dianzi University (Hangzhou Dianzi) "DiTing" student team has developed a "millimeter wave vital signs monitoring system" that can Real-time monitoring of vital signs through snoring, statistics of sleep information of the elderly, and health analysis reports generated in the App. Among them, the main function of the system is to identify apnea syndrome, timely intervention of sleep apnea, and physical examination of hidden diseases of the elderly.
Professor Li Wenjun, Dean of the Electronic Information College of Hangzhou Electronics, introduced that some people snoring very loudly and very long, and they are probably suffering from apnea syndrome. "Snoring is so bad that it's easy to be ignored, and apnea syndrome is actually as deadly as cardiac arrest."
Yue Xueying, the project leader of the team, said that cardiac arrest has received more and more attention because it often occurs in daily life. The apnea syndrome mostly occurs at night when sleeping. Some patients died after falling asleep, but their family members did not know that it was caused by this disease.
According to reports, the terminal of the "millimeter wave vital signs monitoring system" looks like a power bank and can be placed anywhere in the home. Among them, the millimeter wave emitted by the chip in the groove has strong penetrability, can penetrate fog, smoke, dust, etc., and has strong anti-interference ability, which can accurately find the sleeping person at home.
Team members collected the snoring sounds of more than 200 patients in half a year, helping the team successfully crack the "digital code" in the snoring sounds of patients with apnea syndrome. "At a certain distance in a room, millimeter-wave radar must achieve accurate recognition, involving a large number of algorithms." Cheng Siyi, a member of the R&D team, said that we have developed a deep learning-based snoring classification method, which is based on obstructive sleep breathing. The difference between the specific snoring sound produced by the pause hypopnea syndrome (OSAHS) and the ordinary snoring sound, three types of characteristic coefficients of the snoring sound are extracted, and then various types of snoring sound are classified. At the same time, the recognition results of various types of snoring can be used for auxiliary diagnosis of OSAHS.