Title :
Lung sound analysis based methodology to identify asthmatic patient for low power low cost embedded system
Author :
Nitin S. Ambatkar;S. D. Chede
Author_Institution :
Department of Electronics & Telecommunication, Priyadarshini College of Engineering, Nagpur, India
Abstract :
In this paper, lung sounds of healthy normal, asthmatic patients, tuberculosis patients, rheumatoid patients and pneumonia patients are analyzed and classified to diagnose asthmatic patients for designing and development of low power, low cost portable embedded system. These lung sounds are analyzed using wavelet packet transform (WPT) to get different sub-band coefficients. From the sub- bands coefficients of different lung sound signal statistical features vectors are extracted. New LVQ (ANN) is used to categorize lung sound signals as asthmatic or non asthmatic. Data is obtained in normal hospital conditions by a typical stethoscope. Total 26 patient and 10 healthy person databases are tested. The proposed Methodology provides 86.6% of accuracy.
Keywords :
"Lungs","Artificial neural networks","Diseases","Feature extraction","Wavelet packets"
Conference_Titel :
Energy Systems and Applications, 2015 International Conference on
DOI :
10.1109/ICESA.2015.7503448