Combining Multiple speech Recognizers using Voting and Language Model Information

机译：使用投票和语言模型信息组合多个语音识别器

获取原文

页面导航

摘要
著录项
相似文献
相关主题

摘要

In 1997, NIST introduced a voting scheme called ROVER for combining word scripts produced by differnet speech recognizers. This approach has achieved a relative word error reduction of up to 20

机译：1997年，NIST引入了一种名为ROVER的投票方案，用于组合由differnet语音识别器生成的单词脚本。这种方法可以最大减少20个字的相对字错误

著录项

来源
《6th International Conference on Spoken Language Processing ICSLP 2000 Oct.16-Oct.20 2000 Beijing International Convention Center, Beijing, China》|2000年|p.915-918|共4页
会议地点
作者
Holger Schwenk; Jean-Luc Gauvain;
展开▼
作者单位

展开▼
会议组织
原文格式 PDF
正文语种
中图分类世界各国文化与文化事业;
关键词

相似文献

外文文献
中文文献
专利

1. Improving Keyword Recognition of Spoken Queries by Combining Multiple Speech Recognizer's Outputs for Speech-driven WEB Retrieval Task [J] . Masahiko MATSUSHITA, Hiromitsu NISHIZAKI, Takehito UTSURO, IEICE Transactions on Information and Systems . 2005,第3期

机译：通过组合多个语音识别器的输出以执行语音驱动的WEB检索任务，提高口语查询的关键字识别
2. A challenging task in recognizing the speech of the Hearing impaired using normal hearing models in classical Tamil language [J] . Revathi dhanabal, jeya lakshmi Journal of Engineering Research . 2017,第2期

机译：使用古典泰米尔语听力模式识别正常听觉障碍者的语音时的一项艰巨任务
3. A decision-theoretic framework for the evaluation of language models used in speech recognizers [J] . J. R. DELLER Jr., K. H. DESAI, Y. P. YANG Natural language engineering . 2005,第Pt4期

机译：用于评估语音识别器中语言模型的决策理论框架
4. Combining Multiple speech Recognizers using Voting and Language Model Information [C] . Holger Schwenk, Jean-Luc Gauvain International conference on spoken language processing . 2000

机译：使用投票和语言模型信息结合多个语音识别器
5. Handwritten character recognition by combining multiple recognizers using conditional probabilities [D] . Padmanaban, Mageshkumar 2005

机译：通过使用条件概率组合多个识别器来进行手写字符识别
6. Voting-based consensus clustering for combining multiple clusterings of chemical structures [O] . Faisal Saeed, Naomie Salim, Ammar Abdo 2012

机译：基于投票的共识聚类用于组合化学结构的多个聚类
7. Teenagers and young adults are increasingly using social networks as a means to interact and participate in constructing a multiple speech. Companies take direct options with followers in networks and use these virtual structures to approach their target. The purpose of this paper is to study, using empirical and observational methodology, how to build «Coca-Cola» brand image in «Tuenti», followed by the network over the public sector. Among other things, we will see how involved the brand and how followers, what are the issues that are introduced on the inputs and through what kind of formats. In conclusion, we noted that the interest in the brand of free speech to let the followers is just a strategy, the actual entries of «Coca-Cola» are very rare but they all have a high effect, a language that challenges the user to activate and resume his speech so directed. Moreover, there is no mechanism differentiating between information, entertainment and advertising, which combined with continued exposure to advertising impacts across different formats, leads us to propose the need for media education to encourage responsible use critic and social networks by young people. [O] . Carmen Lazo, Marta, Martínez Rodrigo, Estrella, Sánchez Martín, Lourdes 2013

机译：青少年和年轻人越来越多地使用社交网络作为交互和参与构建的手段多个演讲。公司直接选择网络中的关注者并使用这些虚拟结构接近他们的目标。本文的目的是使用经验和观察方法研究如何在«Tuenti»建立«可口可乐»品牌形象，随后是公共部门的网络。除其他外，我们将看到品牌及其追随者的参与程度，以及在投入和投入上引入的问题通过什么样的格式。总之，我们注意到对自由言论品牌的兴趣让追随者只是一个策略，«可口可乐»的实际条目是非常罕见的，但它们都有很高的效果，一种语言，挑战用户如此指示激活和恢复他的演讲。而且，没有机制区分信息，娱乐和广告之间，以及持续接触广告不同形式的影响，使我们建议需要媒体教育来鼓励负责任的使用评论家和年轻人的社交网络。
8. Speech Recognition, Articulatory Feature Detection, and Speech Synthesis in Multiple Languages [R] . Ore, B. M. 2009

机译：语音识别，发音特征检测和多语言语音合成

Combining Multiple speech Recognizers using Voting and Language Model Information

摘要

著录项

相似文献

相关主题

期刊订阅