专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

21. 发明公开

KR1020100021910A 음성인식 방법 및 그 장치 有权
标题翻译：语音识别方法及其设备
公开(公告)号：KR1020100021910A
公开(公告)日：2010-02-26
申请号：KR1020080080573
申请日：2008-08-18
申请人： 엘지전자 주식회사
发明人： 신원호
IPC分类号： G10L15/10 , G10L15/06 , G10L15/08 , G10L15/04
摘要： PURPOSE: A method for speech recognition and an apparatus thereof are provided to improve the voice recognition success rate by adding the pronunciation information of the user to the recognition dictionary. CONSTITUTION: The phonetic information about one or more recognizing subject is respectively matched with the recognition dictionary. In the recognition dictionary, the first recognizing subject is selected(S100). The voice of user is inputted about the first recognizing subject(S110). The voice feature information about the input voice is extracted(S120). The user pronunciation information about the input voice obtains through the acoustic model(S130). The first recognizing subject and user pronunciation information are matched. The matching information is saved in the recognition dictionary(S140).
摘要翻译：目的：提供一种用于语音识别的方法及其装置，通过将用户的发音信息添加到识别词典来提高语音识别成功率。构成：一个或多个识别对象的语音信息分别与识别字典相匹配。在识别字典中，选择第一识别对象（S100）。关于第一识别对象输入用户的声音（S110）。提取关于输入声音的语音特征信息（S120）。通过声学模型获得关于输入声音的用户发音信息（S130）。第一个识别主体和用户发音信息匹配。匹配信息保存在识别字典中（S140）。

22. 发明公开

KR1020040070511A 다기능 디지털 레코더 无效
标题翻译：个人数字记录仪
公开(公告)号：KR1020040070511A
公开(公告)日：2004-08-11
申请号：KR1020030006612
申请日：2003-02-03
申请人： 엘지전자 주식회사
发明人： 신원호 , 은종진 , 김준석
IPC分类号： G10L15/22
摘要： PURPOSE: A PDR(Personal Digital Recorder) is provided to offer a successive voice recognition function. CONSTITUTION: A voice recognition unit(10) receives the voice of a user by a sentence type, recognizes the voice on the basis of a previously composed grammar, and outputs the recognized result as certain character data. A recognition result processing unit(20) recomposes received character data using the character data outputted from the voice recognition unit(10) and the previously composed grammar, and transmits and receives certain data with a previously composed DB according to kinds of the recomposed character data. The recognition result processing unit(20) outputs the result data or the recomposed character data. A DB retrieving unit(40) receives the certain data outputted from the recognition result processing unit(20), and outputs result data corresponding to the received certain data to the recognition result processing unit(20). An output response unit(30) receives the certain data outputted from the recognition result processing unit(20), and outputs certain voice data.
摘要翻译：目的：提供PDR（个人数字录像机），提供连续的语音识别功能。构成：语音识别单元（10）通过句子类型接收用户的语音，基于预先组合的语法识别语音，并将识别结果输出为特定字符数据。识别结果处理单元（20）使用从语音识别单元（10）输出的字符数据和先前组合的语法重构接收到的字符数据，并且根据重构的字符数据的种类与先前组合的DB一起发送和接收某些数据。识别结果处理单元（20）输出结果数据或重组字符数据。 DB检索单元（40）接收从识别结果处理单元（20）输出的特定数据，并将与所接收的特定数据相对应的结果数据输出到识别结果处理单元（20）。输出响应单元（30）接收从识别结果处理单元（20）输出的特定数据，并输出某些语音数据。

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式