DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Lee, Keon Jae | - |
dc.contributor.advisor | 이건재 | - |
dc.contributor.author | Jung, Young Hoon | - |
dc.date.accessioned | 2021-05-12T19:34:25Z | - |
dc.date.available | 2021-05-12T19:34:25Z | - |
dc.date.issued | 2020 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=909975&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/283882 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 신소재공학과, 2020.2,[vi, 69 p. :] | - |
dc.description.abstract | Self-powered flexible acoustic sensors, which mimic the basilar membrane of human cochlear, have attracted significant attention in the field of speaker recognition among many types of acoustic sensors. The flexible piezoelectric acoustic sensors can detect minute speech sound and generate multiple electrical signals, via high sensitivity and multi-electrode structure. The speaker recognition is the process to determine whom say the utterance by using the converted electrical data from original speech sound with machine learning algorithms. The human voice should be accurately detected under background noise conditions since there are many different sounds around the speakers. The speaker recognition under noise conditions have been studied, however, the researches suffer from the absence of sufficient recording data similar to the original utterance. Here, we report the enhancement of flexible PZT piezoelectric acoustic sensor via Nb doping, and its application to speaker recognition under noise condition. After piezoelectric thin film is fabricated by using spin-coating method with Nb-doped PZT sol-gel, highly sensitive piezoelectric acoustic sensor is realized with thin film transferred onto a plastic substrate. Speaker recognition rate under noise conditions is analyzed, in addition to enhanced piezoelectric properties via Nb doping. This paper suggests that the improved flexible piezoelectric acoustic sensor for deep learning algorithms can enable the various artificial intelligence services. | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | Flexible piezoelectric▼aNb-Doping▼aAcoustic sensor▼aMachine learning algorithm▼aSpeaker Recognition | - |
dc.subject | 유연 압전▼aNb 도핑▼a음성 센서▼a딥 러닝▼a화자 인식 | - |
dc.title | Enhancement of flexible PZT piezoelectric acoustic sensor via Nb doping and its application to speaker recognition under noise condition | - |
dc.title.alternative | Nb 도핑을 통한 PZT 유연 압전 음성 센서 성능 향상 및 이를 이용한 잡음 환경에서의 화자 인식 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :신소재공학과, | - |
dc.contributor.alternativeauthor | 정영훈 | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.