Publications
mLung: Privacy-Preserving Naturally Windowed Lung AcLvity DetecLon for Pulmonary PaLents
Published
IEEE-EMBS International Conference on Biomedical and Health Informatics(BHI) and the Body Sensor Networks(BSN) Conferences
Date
2019.05.20
Research Areas
Health
Abstract
mLung is a privacy preserving, energy efficient, naturally windowed, mobile-cloud hybrid pulmonary care service for detecting unusual lung activities like coughing and wheezing from streaming audio and inertial sensor data from a chest-held smartphone for pulmonary patients. mLung employs a combination of: (1) natural windowing of audio data from the patient respiration cycle captured by the inertial sensors, (2) in-phone battery-adaptive speech detection and filtering by a lightweight and energy efficient classifier for patient privacy, and (3) incloud lung and confounding sound classification by a heavyweight and expert supervised classifier. This paper describes the design and implementation of mLung and using novel lung activity data collected by smartphone from 131 patients and healthy subjects, provides empirical evidence that mLung is 15%–25% more accurate in detecting lung sounds when compared to a state-of-the-art phone based internal body sound detection system using specialized microphone hardware.
View publication