2019 · 19 citations · 1 references
Biomedical AcousticsConvolutional Neural NetworkEngineeringMachine LearningWearable TechnologyThoracic UltrasoundAcoustic ModelingSpeech RecognitionCommercial Digital StethoscopeMedical AcousticsData ScienceAudio Signal ProcessingVoice RecognitionAcoustic Signal ProcessingDigital StethoscopeAcoustic CameraComputer ScienceUltrasoundDeep LearningMedical Image ComputingDistant Speech RecognitionDigital AudioLive DemoSpeech ProcessingSpeech InputNon-invasive Digital Stethoscope
We demonstrate a new digital stethoscope system, LungSys, for our users to detect adventitious respiratory sounds automatically. LungSys includes a commercial digital stethoscope and a software application installed on an Android mobile tablet. The digital stethoscope converts an acoustic sound from the users' chest to electronic signals and transmits the signals to a mobile tablet through a built-in Bluetooth device. Our custom software application in the tablet provides a real-time analysis of the lung sound using our proposed neural network model bi-ResNet(BRN) and identifies any adventitious respiratory sound to users. Since LungSys is based on a non-invasive digital stethoscope and our proprietary deep learning algorithm, it allows users who do not have any professional skill to perform respiratory diagnosis conveniently.
1