Auscultating Diagnosis Support System by Using Self-Organizing Map II -Classification of Eustachian tube insufflation sound-
Yutaka Suzuki, Osamu Sakata, Shun'ichi Imamura, Shuichiro Endo, Keisuke Masuyama, Masayuki Morisawa · 2019
The Eustachian tube insufflation method is widely used for diagnosis and medical treatment. However, because the Eustachian tube insufflation method is auscultated by only the doctor who enforces it, the Eustachian tube insufflation sound is not left behind as data, and diagnosis may change depending on the doctors because the sound on which the diagnosis is based is subjective. Therefore, this research considered using feature extraction for classifying the difference in the Eustachian tube insufflation sound according to the disease condition using a self-organizing map (SOM) for the purpose of improving the objectivity and reliability of diagnosis by quantification of the Eustachian tube insufflation method. In this study, various feature vectors to be input to the SOM were prepared and examined, and the result of refinement concerning the optimal condition is reported.