No abstract

Abstract.The air flow during the respiration process produces lung sound and provide information on lung health. Automatic lung sound recognition becomes one of the areas of interest to researchers in the field of biomedical signal processing. Signal complexity measurement becomes one of features extraction method for lung sound analysis. Some signal complexity measurement technique that is often used for example are entropy, fractal dimension, and high-order statistics. In this study conducted multiscale Hjorth descriptor measurements on lung sounds for lung sound classification. The results showed that the complexity on a scale of 1-5 yield 95.06% accuracy. Multiscale analysis succeeded in improving the accuracy of the lung sound classification. The higher the scale that is used does not guarantee to increase the accuracy.

No abstract

No abstract

No abstract