Dog cough sound classification using artificial neural network and the selected relevant features from discrete wavelet transform
Conference proceedings article
Authors/Editors
Strategic Research Themes
No matching items found.
Publication Details
Author list: Kakabutr P., Chen K.S., Wangvisavawit V., Padungweang P., Rojanapornpun O.
Publisher: Hindawi
Publication year: 2017
Start page: 121
End page: 125
Number of pages: 5
ISBN: 9781467390774
ISSN: 0146-9428
eISSN: 1745-4557
Languages: English-Great Britain (EN-GB)
Abstract
Coughing is one of the important signs of several diseases in dogs. There are two types of dog cough: dry cough and productive cough. The latter is most often associated with an infectious condition. It is difficult to differentiate between the two types even by experienced practitioners. In this paper, an automatic cough sound classification using neural network is introduced. A discrete wavelet transform is employed to decompose the cough sound into low frequency and high frequency components. The statistical features of these components are used as the sound features. The discrimination power of these features in classification are evaluated. Finally, the artificial neural network is used to classify dog cough sound using a subset of discriminant features. The experimental results show that classifying dog cough sounds needs only one fourth of all features and an average accuracy as high as 90% is achievable. ฉ 2017 IEEE.
Keywords
Dog cough sound
Contact Information