Lung Sound Classification using Wavelet Transform and Entropy to Detect Lung Abnormality

Abstract

Lung sounds contain necessary information about the health of the lungs and respiratory tract. They have unique and distinguishable pattern associated with the abnormalities probably occurred in the lungs or in the respiratory tract. Many researches have attempted to develop variety of methods to classify the lung sounds automatically. Of the methods, wavelet transform is one frequently used for physiological signals analysis. Commonly, the use of wavelet in feature extraction is used to break down the lung sounds into several sub-bands prior to the calculation of some parameters. In this research, five classes of lung sound that were obtained from various sources were used. Then, the wavelet analysis process were carried out using Discrete Wavelet Transform (DWT) and Wavelet Package Decomposition (WPD) analysis and entropy calculation as feature extraction. In DWT process, the highest accuracy obtained was 97.98% using Permutation Entropy (PE), Renyi Entropy (RE) and Spectral Entropy (SEN). Whereas in WPD, the highest accuracy produced is 98.99% by using eight sub-bands and RE. The results in this research are quite competitive if compared with the results of previous studies that used the wavelet method with the same datasets.