PC based multichannel electronic stethoscope designing and materialize

In this study, a PC based multi-channel data acquisition and processing systemhas been developed in order to record and analyze the lung sounds in real time.The signals were taken from the chest wall by multi-channel. This systemconsists of a specially designed sound acquisition system, data acquisitionsystem (Instrunet model) and graphic user program interfaces for computerrecords and analysis. The lung sounds are acquired from 14 different on lunglocated on the chest wall and the trachea. Thus the local analysis has beenrealized by differentiating the lung sound listening activity to 14 different lungsites.The different time and frequency domain analysis of the lung sounds acquiredas multi-channel have been conducted and the results have been observed. Inthe time domain, time amplitude waveform, time expanded waveform, RMSvalues, frequency changes cohesive to zero crossing velocity, duration andamplitude of inspiration and expiration, wheeze sounds detection, cracklesounds detection and different duration and amplitude resolutions wereobserved. The spectral analysis of the lung sound signals in the frequencydomain have been performed by predicting the power density spectrums withFast Fourier Transform method.Afterwards the classification of lung sounds has been performed by using thefeature vectors obtained from the spectral analysis with the help of multilayerperception neural networks.In conclusion it has been proved that with the analysis of visual-acousticpresentations and signal processing methods of the lung sounds, moreinformation concerning the lung diseases diagnosis may be obtained with anoninvasive method.