Evaluation of abnormal respiratory sound "crackle" using time-expanded waveform and FFT analysis

Issue title: ICMIT '01

Article type: Research Article

Authors: Kawamura, Takeo^{a; *} | Matsumoto, Tsuneo^a | Miura, Gouji^a | Matsunaga, Naofumi^a | Kido, Shoji^b | Jiang, Zhongwei^c

Affiliations: [a] Department of Radiology, Yamaguchi University School of Medicine, Yamaguchi 755-8505, Japan | [b] Department of Computer Science and Systems Engineering, Yamaguchi University, Japan | [c] Department of Mechanical Engineering, Yamaguchi University, Japan

Correspondence: [*] Corresponding author. Tel.: +81 836 22 2283; Fax: +81 836 22 2285; E-mail: kawa@mx52.tiki.ne.jp

Abstract: Purpose: Many diseases such as bronchial asthma or pneumonia are detected by chest auscultation with a stethoscope. However, there are no obvious criteria for evaluating the chest auscultation findings. In the present study, chest auscultation findings were evaluated by digital computer analysis. Material and method: "Crackle" was evaluated among abnormal respiratory sounds. Crackles with very short duration were extracted from recorded respiratory sounds, and FFT (Fast Fourier Transform) analysis was used. Result: Spike-like waves could be recognized visually in the time-amplitude plot of respiratory sounds. The durations of each crackle ranged between 12.6 and 31.3 msec. The high/low-pitched sound ratio calculated by FFT analysis ranged between 0.059 and 3.24. Conclusion: The characteristics of crackles could be evaluated objectively by computer analysis. The high/low-pitched sound ratio correlated with the computed tomography (CT) findings and the ratio is useful for differential diagnosis of the disease.

DOI: 10.3233/JAE-2003-270

Journal: International Journal of Applied Electromagnetics and Mechanics, vol. 18, no. 1-3, pp. 11-21, 2003