Huey-Dong Wu

WHEEZE RECOGNITION BASED ON 2D BILATERAL FILTERING OF SPECTROGRAM

This paper describes the design of a low-cost and high performance wheeze recognition system. First, respiratory sounds are captured, amplified and filtered by an analog circuit; then digitized through a PC soundcard, and recorded in accordance with the Computerized Respiratory Sound Analysis (CORSA) standards. Since the proposed wheeze detection algorithm is based on the spectrogram processing of respiratory sounds, spectrograms generated from recorded sounds have to pass through a 2D bilateral filter for edge-preserving smoothing. Finally, the processed spectra go through an edge detection procedure to recognize wheeze sounds. Experiment results show a high sensitivity of 0.967 and a specificity of 0.909 in qualitative analysis of wheeze recognition. Due to its high efficiency, great performance and easy-to-implement features, this wheeze recognition system could be of interest in the clinical monitoring of asthma patients and the study of physiological mechanisms in the respiratory airways.

Wheeze Detection Using Cepstral Analysis in Gaussian Mixture Models

Traditional wheezes detection methods are based on the frequency and durations of acoustic signal or the location of peaks from successive spectra. In these methods, the discriminative threshold used to identify peaks usually is fixed empirically. Therefore, accuracy of detected wheeze is affected by environment noise and artificial factors. The objective of this study is to classify normal and abnormal (wheezing) respiratory sounds using cepstral analysis in Gaussian Mixture Models. The sound signal is divided in overlapped segments, which are characterized by a reduced dimension feature vectors using Mel-Frequency cepstral coefficients. In this study the speakeris wheeze. During the test phase, an unknown sound is compared to all the GMM models and the classification decision is based on the Maximum Likelihood criterion. In these processes, identification is based on threshold value. If the threshold is bigger than zero, the sound is normal. Otherwise, the sound is wheeze. From experimental results, when the Gaussian mix number is 16, the accuracy of identification of wheeze is up to 90%.

Respiratory Wheeze Detection System

Respiratory sound is associated with many lung diseases. By observing respiratory sound symptoms, we can know more about lung conditions. In this research, we construct an efficient lung sound recording system according to CORSA, and develop a spectrogram process flow technique to object wheeze. It is a low cost and efficient system. In clinic test, we also can precisely objective wheeze up to about 89%

Ambient noise canceller in pulmonary sound using WHT transform domain adaptive filter

In the process of signal processing of pulmonary sounds, the cancellation of ambient noise is very important. Because the ambient noise is a wide band signal in the frequency domain, it usually uses an adaptive noise canceller (ANC) structure to cancel the ambient noise. However, the usually used algorithm, normalized least mean square (nLMS) algorithm, may fail when dealing with the ambient noise due to the time-varying system because of its low convergence speed. We use a transform domain adaptive filter (TDAF) with a Walsh-Hadamard transform (WHT) to improve the convergence speed. The simulation results show that this structure would cancel the ambient noise more efficiently.

Automatic Wheezing Detection Based on Signal Processing of Spectrogram and Back-Propagation Neural Network

Wheezing is a common clinical symptom in patients with obstructive pulmonary diseases such as asthma. Automatic wheezing detection offers an objective and accurate means for identifying wheezing lung sounds, helping physicians in the diagnosis, long-term auscultation, and analysis of a patient with obstructive pulmonary disease. This paper describes the design of a fast and high-performance wheeze recognition system. A wheezing detection algorithm based on the order truncate average method and a back-propagation neural network (BPNN) is proposed. Some features are extracted from processed spectra to train a BPNN, and subsequently, test samples are analyzed by the trained BPNN to determine whether they are wheezing sounds. The respiratory sounds of 58 volunteers (32 asthmatic and 26 healthy adults) were recorded for training and testing. Experimental results of a qualitative analysis of wheeze recognition showed a high sensitivity of 0.946 and a high specificity of 1.0.

Combination of frequency and amplitude-modulated model for the synthesis of normal and wheezing sounds

Based on communication theory, this study proposes a model to synthesize normal and wheezing sounds. The model included five parts: the flow source as a transmitter, the frequency and amplitude-modulated (FM–AM) sounds, the accompanying noise as a modulator, the airway wall as a medium, and the microphone as a receiver. The hypothesis of modulation builds on that the deviation of frequency and amplitude of the sounds which cause from the deviation of collision speed of the air flow on the wall. The model was successful to simulate the normal breath and wheezing sounds. Furthermore, it provided a correct proof for the CORSA description, which indicates that the wheeze was contained in the domain frequency at 400xa0Hz, but a number of investigators have suggested that the range is actually between 80–1,600xa0Hz and 350–950xa0Hz by filter theory. This study modifies the signal source in Wodicka et al. model, and describes it in functional blocks. In fact, the design of the signal source base on the knowledge of the lung sound studies, especially the analysis of components in the frequency and time domains. We synthesized the required components to reproduce the lung sounds, and proposed a mechanism of wheeze which was examined by the computer simulation in the points of the system engineering view.

Real-time mobile-to-mobile stethoscope for distant healthcare

This study presented to insert a small-scaled microphone of the ear-set of a smart mobile (Amazing A6, Taiwan MobileTM) into an eartip of the stethoscope, the sound can be transmitted by the mobile when it dialed to another phone or mobile. In this study, we employed another smart phone (Galaxy R GT19103, SamsungTM) to be the receiver. The results were showed by spectrogram which demonstrated the components of frequencies of the sounds. Finally, we proposed to improve the study by database.

The feasibility study of mobile-to-mobile communication for auscultation of heart sound and lung sound

This study presented to insert a small-scaled microphone of the ear-set of a smart mobile (Galaxy Beam, GT-I8530, SamsungTM) into an eartip of the stethoscope, the sound can be transmitted by the mobile when it dialed to another phone or mobile. In this study, we employed another smart phone (Windows Phone 8X, hTC TM) to be the receiver. The results were showed by time sequence signals as well as spectrograms which demonstrated the components of frequencies of the sounds. Surprisingly, the recorded sounds of subjects mobile were much noisier than those of doctors, because of the very advanced technology of mobile communication which included the auto gain control (AGC), digital filter, noise cancellation and amplifier technologies to improve the quality of communication. We believe that the mobile-to-mobile communication is feasible to be the mean of distant auscultation. The perspective on the filtering modes, we got the conclusions of the distant auscultation, i.e. lung sound is better than heart sound.

Using Back-Propagation Neural Network for Automatic Wheezing Detection

This study describes the design of a fast and high performance wheeze recognition system. The proposed wheezing detection algorithm is based on order truncate average (OTA) and back-propagation neural network (BPNN). Some features are then extracted from the processed spectra to train a BPNN. Eventually, the new testing samples go through the trained BPNN to recognize whether they are wheezing sounds. Experimental results show a high sensitivity of 0.946 and a specificity of 1.0 in qualitative analysis of wheeze recognition.

Distant auscultation system for detecting lung sounds of patients on ambulances

The sound of the siren of the ambulance is for the safety of the road transportation, but interferes the auscultation of the lung and heart sounds. The system implementation is included by (1) ACER Aspire 17 notebook as a server in right side; (2) HwaWei Amazing A6 smart mobile as a hot point in the middle; and (3) ACER Aspire 5 notebook as a client which the settings include IP address of host computer, and client, read and, write privileges of the data sockets, and running of NI data socket manager, and data socket server. Therefore, the parameters in the real-time DAS are verified as the better ones to prepare for the services on the ambulance.