Development of digital stethoscope system making use of patients at home

114例患者获得知情同意，并对其呼吸音和CT图像进行保存。用数字听诊器（Power听诊器，Starky Japan, Yokohama）记录胸部12个部位的呼吸声数据。在吸气期和呼气期进行胸部高分辨率CT （HRCT）检查。首先，开发了用小波变换从呼吸声数据中检测异常域的程序。其次，为了详细评价裂纹的特征，我们使用了持续时间初始挠曲宽度（IDW）和两周持续时间（2CD）三个参数。开发了细裂纹与粗裂纹的判别程序。这些参数与HRCT具有良好的相关性。并且，为了对裂纹特性进行另一种评价，使用了相邻波形区间的统计数据。实现了波形时间间隔统计值的获取、直方图的生成和相似度的识别。使用相邻波形间隔比值的统计值识别出医生听诊视图中97%的粗裂声、80%的粗裂声和98%的正常声。针对用于呼吸声音采集的数字听诊器，本文做了无线电子听诊分析系统作为实验。它大致分为听诊器模块、无线音频传输和接收模块以及声音分析软件。如果需要，它也可以用于组合模块，商业电子听诊器，外部存储等。

Informed consent was obtained to the patient of 114 examples, and preservation of their respiratory sounds and CT images were performed. The respiratory sound data were recorded with a digital stethoscope (Power Stethoscope, Starky Japan, Yokohama) from 12 sites on the chest. Chest high-resolution CT (HRCT) at the inspiratory and expiration phase was performed.First, the program, which detects an unusual domain from the respiratory sound data with wavelet conversion, was developed. Secondary, to evaluate the characteristics of the crackle in detail, we used three parameters of the duration initial deflection width (IDW) and two-cycle duration (2CD). The distinction program of fine crackle and coarse crackle was developed. Good correlation was obtained between these parameters and the HRCT. And, for another evaluation of the characteristics of the crackle, statistics of the adjoining waveform interval were used. The procedure of acquisition of wave-like time interval statistics value, creation of histogram and discernment by the degree of similar performed the technique. And the statistics value using the ratio of the adjoining waveform interval identified 97% of the coarse crackles of a doctor's auscultation view, 80% of a coarse crackle, and 98% of normal sound.About the digital stethoscope for respiratory sound collection, the wireless electronic auscultation analysis system was made as an experiment this time. It is divided roughly into a stethoscope module, wireless audio transmission and a receiving module, and sound analysis software. It also becomes possible to use to combine a module if needed, a commercial electronic stethoscope, an external storage, etc.