engleski

Asthmatic Wheezes Detection - What Contributes the Most to the Role of MFCC in Classifiers Accuracy?

Asthma is one of the top five chronic diseases globally and the most common chronic disease among children. It is the most likely cause of recurrent wheezing in children, so computerized respiratory sound analysis is an important diagnostic aid. This research compares the efficiency of the classification algorithms applied both on signals available on the internet and signals recorded on children in real-life clinical settings. The paper proves that the features with logarithmic distribution of energy filter bank along the frequency domain embedded in MFCC, result in better wheezes recognition in an auscultatory breathing signal than spectral features and the similar energy filter bank features which do not have logarithmic distribution along the frequency domain. Furthermore, the paper demonstrates that the SVM classifier performs better than other classifiers applied on signals acquired under ideal and suboptimal conditions.

asthma; classification; MFCC; SVM

nije evidentirano