東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Prosody dependent speech recognition...

Chen, Ken.

Linked to FindBook

Google Book

Amazon

博客來

Prosody dependent speech recognition on American radio news speech.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Prosody dependent speech recognition on American radio news speech./
Author:	Chen, Ken.
Description:	124 p.
Notes:	Source: Dissertation Abstracts International, Volume: 65-11, Section: B, page: 5917.
Contained By:	Dissertation Abstracts International65-11B.
Subject:	Engineering, Electronics and Electrical. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3153265
ISBN:	0496137549

Prosody dependent speech recognition on American radio news speech.
Chen, Ken.

Prosody dependent speech recognition on American radio news speech. - 124 p.

Source: Dissertation Abstracts International, Volume: 65-11, Section: B, page: 5917.

Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2004.

Prosody (the melody and rhythm of natural speech), although important for human speech recognition, has not been fully utilized in large vocabulary continuous speech recognition. In this dissertation, we propose a novel "prosody-dependent speech recognition" framework, in which word and prosody are recognized simultaneously for the purpose of improving word recognition accuracy. We review the linguistic literature on how prosody is used in human speech communication, what the structure and function of prosody is, and how prosody affects acoustic realization of the segmental and suprasegmental units of speech. We conduct information-theoretic analysis, proving that when prosody is modeled, better word recognition can be achieved through the interaction between the acoustic model and the language model. We conduct detailed experiments to determine the set of allophonic HMMs or probability distributions that are sensitive to prosody, under the guidance of linguistic prior knowledge and empirical selection rules. We measure the effects of prosody on the language modeling and on the pronunciation modeling, and propose a factored approach, which leverages the strong dependence of prosody over syntax, to improve the robustness of the N-gram language modeling. We also develop an automatic prosody labeling system as a way to reduce the human labeling cost based on a simplified version of the Tones and Break Indices system. In the word recognition experiment on the Boston University Radio News Corpus, we find that our system is able to reduce word error rate by as much as an absolute 11%, as compared with a conventional prosody-independent HMM-based automatic speech recognizer that has comparable parameter count. Our research confirms that explicit modeling of prosody in HMM based automatic speech recognizers can improve word recognition on American Radio News speech.

ISBN: 0496137549Subjects--Topical Terms:

626636
Engineering, Electronics and Electrical.

Prosody dependent speech recognition on American radio news speech.
LDR:02819nmm 2200289 4500 001 1850469
005 20051208095324.5
008 130614s2004 eng d
020 $a 0496137549
035 $a (UnM)AAI3153265
035 $a AAI3153265
040 $a UnM $c UnM
100 1 $a Chen, Ken. $3 1938393
245 1 0 $a Prosody dependent speech recognition on American radio news speech.
300 $a 124 p.
500 $a Source: Dissertation Abstracts International, Volume: 65-11, Section: B, page: 5917.
500 $a Adviser: Mark Hasegawa-Johnson.
502 $a Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2004.
520 $a Prosody (the melody and rhythm of natural speech), although important for human speech recognition, has not been fully utilized in large vocabulary continuous speech recognition. In this dissertation, we propose a novel "prosody-dependent speech recognition" framework, in which word and prosody are recognized simultaneously for the purpose of improving word recognition accuracy. We review the linguistic literature on how prosody is used in human speech communication, what the structure and function of prosody is, and how prosody affects acoustic realization of the segmental and suprasegmental units of speech. We conduct information-theoretic analysis, proving that when prosody is modeled, better word recognition can be achieved through the interaction between the acoustic model and the language model. We conduct detailed experiments to determine the set of allophonic HMMs or probability distributions that are sensitive to prosody, under the guidance of linguistic prior knowledge and empirical selection rules. We measure the effects of prosody on the language modeling and on the pronunciation modeling, and propose a factored approach, which leverages the strong dependence of prosody over syntax, to improve the robustness of the N-gram language modeling. We also develop an automatic prosody labeling system as a way to reduce the human labeling cost based on a simplified version of the Tones and Break Indices system. In the word recognition experiment on the Boston University Radio News Corpus, we find that our system is able to reduce word error rate by as much as an absolute 11%, as compared with a conventional prosody-independent HMM-based automatic speech recognizer that has comparable parameter count. Our research confirms that explicit modeling of prosody in HMM based automatic speech recognizers can improve word recognition on American Radio News speech.
590 $a School code: 0090.
650 4 $a Engineering, Electronics and Electrical. $3 626636
650 4 $a Language, Linguistics. $3 1018079
650 4 $a Computer Science. $3 626642
690 $a 0544
690 $a 0290
690 $a 0984
710 2 0 $a University of Illinois at Urbana-Champaign. $3 626646
773 0 $t Dissertation Abstracts International $g 65-11B.
790 1 0 $a Hasegawa-Johnson, Mark, $e advisor
790 $a 0090
791 $a Ph.D.
792 $a 2004
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3153265