An Embedded Classifier of Lung Sounds based on the Wavelet Packet Transform and ANN

Abstract

Electronic auscultation is an efficient technique for capturing lung sounds and analysis of these sounds permits to evaluate the health condition of the respiratory system. Since the lung sound signals are non-stationary, the conventional method of frequency analysis is not highly successful in diagnostic classification. This paper presents a method of analysis of lung sound signals using the Wavelet Packet Transform (WPT), and classification using artificial neural network (ANN). Lung sound signals were decomposed into the frequency sub-bands using the WPT and a set of statistical features was extracted from the sub-bands to represent the distribution of wavelet coefficients. An ANN is trained and then applied to classify the lung sounds into one of three categories: normal, wheeze or crackle. This classifier was embedded in a microcontroller to provide an automated and portable device that will provide support for research and diagnosis in the evaluation of the respiratory function.