Wheeze Recognition Based on 2d Bilateral Filtering of Spectrogram

Abstract

ABSTRACT

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

“…By the way, the 8 KHz sampling rate of telephone and VoIP has been evaluated to record wheezing sound well. The sampling rate of the study was identified and supported the results of our previous studies [8][9][10][11].…”

“…By the way, the 8 KHz sampling rate of telephone and VoIP has been evaluated to record wheezing sound well. The sampling rate of the study was identified and supported the results of our previous studies [8][9][10][11]. In addition the reasons of noises in physiologic signals are usually band-limited, therefore, the optimal sampling frequency can be used to simplify the system design.…”

“…The sampling rates of the whole communication processes were listed as follows: The electronic stethoscope was 4 KHz of sampling rate, and the mobile communication was 8 KHz of sampling rate. In our previous studies [8][9][10][11], the theoretical proof was performed by Fourier transform, as well as the statistical analysis was operated by computer simulation to achieve the box plot of the root-mean-square (RMS) errors from 100 trials for each variable. The sampling rates were 44.1(Group 1), 22.05(Group 2), 11.025(Group 3), 5.5125(Group 4), and 2.76125(Group 5) KHz.…”

“…However, the real heart, and lung sounds are not easy to be defined, so we take the reliability of the well-known brand of the 3MTM Littmann® , and the modern Bluetooth® data transfer technology to define the real sound from chestpiece. The spectrogram is a convenient tool for speech recognition [7,8]. The same applications can be employed in the respiration acoustic signals.…”

Spectrogram Enhancement By Edge Detection Approach Applied To Bioacoustics Calls Classification

“…If the smoothing kernel is quite large, the detrimental effect is reduction in the detection accuracy, especially at low SNR. Meanwhile, the authors in [24] passed the spectrogram through 2D bilateral filter to reduce noise and preserve its patterns edges. The filtered spectrogram is sent to two parallel processing paths, at which the first path extracts significant patterns from background noise.…”