Classification of Respiratory Sounds in Smokers and Non-smokers using k-NN Classifier
Respiratory sounds contain significant information on physiology and pathology of the lung and the airways. Its analysis provides vital information of the present condition of the lungs. Pulmonary disease is a major cause of ill-health throughout the worl
- PDF / 849,655 Bytes
- 6 Pages / 595 x 791 pts Page_size
- 15 Downloads / 207 Views
1 School of Mechatronic Engineering, Universiti Malaysia Perlis (UniMAP), Kampus Pauh Putra, Perlis. Malaysia. Faculty of Electronics and Computer Engineering, Universiti Teknikal Malaysia Melaka (UTeM), Hang Tuah Jaya, Melaka, Malaysia. 3 School of Electronics Engineering (SENSE), VIT University, Vellore, India.
. Abstract— Respiratory sounds contain significant information on physiology and pathology of the lung and the airways. Its analysis provides vital information of the present condition of the lungs. Pulmonary disease is a major cause of ill-health throughout the world. The frequency spectrum and the amplitude of sound, i.e. tracheal or lung sounds without adventitious sound components (wheeze), may reflect airway dimension and other their pathologic changes (airways obstruction). Wheezes may have acoustic features indicating not only the presence of abnormality in the respiratory system but also the severity and locations of airway obstruction most frequently found in asthma and also found in smoker but not all smokers have airway obstruction. The significance of this study is to develop a classification system to distinguish between normal and smoker from respiratory sounds. 15 smokers and 15 non-smokers are recruited to collect respiratory sounds using Wireless Digital Stethoscope. The performance analysis of the K-Nearest Neighbor (k-NN) classifier, which uses entropy as the suitable feature, revealed that the classification accuracy on non-smokers and smokers are 89.33% and 78.67% respectively.
Keywords— Respiratory sounds, Airways Obstruction, Fourier Transform, K-Nearest Neighbor
I.
INTRODUCTION
Lung sounds are part of the sound of breathing or commonly called respiratory sound. Respiratory sounds include the sounds from the mouth, trachea, and the lung sounds occur in the chest (chest wall) [1]. As informed before, abnormal breath sounds like wheeze and rhonchi at times manifest similar morphologies and pathological features of lung airways obstruction that occur in smokers [2]. This may pose problems to proper diagnosis and evaluation of the underlying respiratory condition by human auscultation. Based on the spectral features (frequency domain), it is able to detect and classify normal and abnormal breath sounds in smokers using WISE Digital Stethoscope. The significance of this study is to develop a classification system using K-Nearest Neighbor (k-NN) classifier to distinguish between smoker and non-smoker from respiratory sounds.
F. Related works In recent years, researchers are developing and proposing a numerous method to classify lung sounds into two categories: normal and abnormal respiratory sounds (RS). As the main focus of the research, the process must be independently of different breath sounds signal of the subjects. In early 1980’s, respiratory pathologies were recognized using computerized respiratory sound analysis [3]. This section will discuss a few related works that have been done by other researchers in recent years about the computerized respiratory sound analysis. A. R
Data Loading...
