Combined Use of Speech and Bronchial Breath Sounds Acquired by Stethoscope for Person Identification

• Main Authors: Xin-Ying Li, 李欣穎
• Other Authors: Wei-Ho Tsai
• Format: Others
• Language: en_US
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/kkcnbx

碩士 === 國立臺北科技大學 === 電子工程系 === 106 === This work aims to identify patients by their body sounds acquired with stethoscopes. Bronchial breath sounds are one type of body sounds, which can be picked up by a stethoscope. They are formed by glottic or tracheal or main bronchi and are inhaled or exhaled through the mouth and nose. Individuals produce distinctive bronchial breath sounds, which therefore can be used for identifying and distinguishing patients. On the other hand, considering that stethoscopes can also capture the patient’s voice, and thus both bronchial sounds and human voices can be used together to perform patient identification, when the equipment is placed in the bronchial position. In this study, respiratory and articulation sounds of the neck bronchi were collected. By calculating their Mel-frequency cepstral coefficients and using the Gaussian mixture model to determine the identity, we further used “I-Vector & LDA” method to improve the accuracy of identity recognition. This method can achieve a recognition accuracy of 96.87% under the database test of 16 subjects.