객체데이터베이스를 이용한 화자인식처리 기법의 사례연구

다국어 초록 (Multilingual Abstract)

This Paper doesn’t aim at disposal for Phonetic Recognition. It feel it’s way for divide speech applying not recognition system but Jasmine OODBMS which extracts only the character of phonetic and applies with real object database. For this, I intend to set up the theses of deportation way of phonetic data. Chaos theory which intend to use natural phenomena has studied, it can contact to irregular time serial signal. As analyzing dynamic information which is hided in complicate change of signal through Chaos theory, a lot of studying which intend to standardize a
signal is going on. At this Paper, To deportate character of phonetic, I plan to apply Chaos to Phonetic Recognition. Phonetic is came out fluctuation in a wave frequency and change in time.
So let like this deportated phonetic data store with Jasmine database, real OODBMS. It is able to use with a parameter to divide a speech. Jasmine database which is objectdirection database holding wild multimedia class is able to optimize suitably in web. This Paper will apply this to practice through embodying Speech Recognition in web using OODBMS.
-

목차 (Table of Contents)

• Ⅰ. 서론 1
• Ⅱ. 본론 2
• 1. 클라이언트-서버 데이터베이스 시스템 방법 2
• 2. 관계형 데이터베이스 3
• 3. 객체지향 데이터베이스 관리 시스템 (OODBMS) 4
• 4. 자스민 객체데이터베이스 5
• 5. 음성신호처리 8
• 가. 소리형식 8
• 나. 소리테이터의 유형 9
• 다. 음성특징 축출 과정 9
• 라. LPC Analysis 9
• 마. Bank-of-filters analysis 9
• 바. HTK 10
• 사. Lyapunov 차원을 이용한 음성처리 10
• Ⅲ. 결론 11