CASIA OpenIR  > 毕业生  > 硕士学位论文
Alternative TitleThe Research and Application of Domain Limited Continuous Speech Recognition and Key Words Spotting
Thesis Advisor李成荣
Degree Grantor中国科学院研究生院
Place of Conferral中国科学院自动化研究所
Degree Discipline模式识别与智能系统
Keyword连续语音识别 关键词检测 人机对话系统 儿童语音识别 Continuous Speech Recognition Key Words Spotting Human Machine Dialog System Children Speech Recognition.
Abstract语音识别技术是人类迈向高度智能化和自动化的信息社会所必备的关键技术之一。经过几十年的艰苦探索和研究,语音识别技术研究获得了极大的发展,某些比较成熟的技术已经逐步应用于日常生活中。以命令词识别为代表的中小词汇量语音识别技术是近几年嵌入式语音技术应用的主要方向,在实用中取得比较大的进展。随着信息时代的飞速发展,它们已经不能满足用户的实际需要,限定领域的连续语音识别技术和关键词检测技术是语音识别技术实用化研究中相当重要的发展方向。但它们在研究和实用化方面的难度都比较大,具有重要的研究意义和实用价值。 本文开展的工作主要集中于限定领域的连续语音识别技术与关键词检测技术的研究和应用,以及儿童语音识别的初步探索。概括起来有以下几个方面:1.在已有的连续语音识别引擎的基础上,研究开发了适于新型智能娱乐机器人使用的限定领域的连续语音识别引擎。在智能机器人系统中应用表明,该识别引擎的性能优越。2.在连续语音识别引擎的基础上,开发了基于连续语音识别的关键词检测系统,该系统已经应用于人机对话聊天系统。3.对儿童语音识别技术做了初步的探讨。总结了儿童语音的定义、儿童的声道特点以及儿童的说话特点,然后介绍了我们采集的儿童语音数据库,最后给出了提高儿童语音识别率的一些方法。
Other AbstractWith the development of science and technology, Speech Recognition becomes one of the key techniques for human to realize intelligent and robotic information society. After decades of laborious exploration and research, Speech Recognition technology has gained significant progress. Some techniques in this field are being applied to human’s daily life gradually. Medium and small vocabulary speech recognition techniques, especially command words recognition, being main directions for application of embedded speech technology recently, have gained great achievements in real use. As the development of information, they can’t satisfy the demands to modern society. Continuous Speech Recognition and Key Words Spotting(KWS) are very important directions in the research and application of Speech Recognition technology. Though very difficult, they have significant importance in both research and application. Domain Limited Continuous Speech Recognition and Key Words Spotting, together with a primary study on Children Speech Recognition, had been studied in this paper. The research can be summed as follows: 1.Based on the former Continuous Speech Recognition technology, the researchers have developed a Domain Limited Continuous Speech Recognition engine, which aimed at the application on new type intelligent entertainment robot. The result showed that the engine had very good performance in the application on intelligent robot system. 2.Continuous Speech Recognition based Key Words Spotting system on the basis of Continuous Speech Recognition engine have been developed by the researchers. And the system had been used in human machine dialog system. 3.The researchers also have made a primary study on Children Speech Recognition techniques. This paper presented the definition of children speech, characteristics of children vocal tracts, characters of children speech, and the children speech database we developed. Finally, the researchers proposed some techniques for improving the recognition rate of children speech.
Other Identifier200228014603538
Document Type学位论文
Recommended Citation
GB/T 7714
陈伟. 限定领域连续语音识别与关键词检测的研究与应用[D]. 中国科学院自动化研究所. 中国科学院研究生院,2005.
Files in This Item:
There are no files associated with this item.
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[陈伟]'s Articles
Baidu academic
Similar articles in Baidu academic
[陈伟]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[陈伟]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.