Multi-resolution decomposition applied to crackle detection

Abstract

Crackles, heard over the lungs in a variety of diseases, are one of the most important physical signs in clinical medicine. They have an explosive pattern in the time domain, with a rapid onset and short duration. The timing, repeatability and shape of crackles are important parameters for diagnosis. Therefore, automatic detection of crackles and their classification as fine and coarse crackles have important clinical value. Since the multi-resolution decomposition technique can give high resolution in both time and frequency, it can be exploited to detect crackles and to classify them according to the information in each scale. In this paper, we present a new method for crackle detection based on the continuous wavelet transform. The theory, methods and experimental results are given in detail in this paper.