This paper addresses the problem of speech recognition under telephone channel conditions using data simulation method and HMM(Hidden Markov Model)adaptation.
该文研究了基于数据模拟方法和HMM（隐马尔科夫模型）自适应的电话信道条件下语音识别问题。
This paper describes the use of multi-layer perception model of neural network in speech recognition.
本文研究神经网络的多层感知器模型在语音识别中的应用。
In order to solve these problems, we proposed a single feature vector recognition model based on whole time-frequency information structure of digit speech.
为了解决这些问题，我们提出了基于数字语音时频信息整体结构的单特征向量识别模型。
A framework model of independent speech recognition system based on the flexible and extensible architecture is put forward, and some correlative theories are introduced.
提出了一个柔性可扩展体系结构非特定人语音识别系统的框架模型，介绍了相关实现原理。
In this paper, a Chinese isolated word recognition system is established based on the source-filter generation model combined with the acoustic characteristics of whispered speech.
本文根据耳语音信号发音模型，结合耳语音的声学特性，建立了一个汉语耳语音孤立字识别系统。
Furthermore it is very important that we use language model, syntax and accidence model in middle or big glossary continuous speech recognition.
而语言模型、语法及词法模型在中、大词汇量连续语音识别中是非常重要的。
This paper also introduces HMM model and speech digital signal processing associated with speech recognition .
理论上详细介绍了HMM模型及与语音识别相关的语音数字信号处理。
Human-ma-chine interlocution of Smartpanda System is implemented by means of speaker independent continuous speech recognition technology and dialogue model.
该系统利用大词汇量非特定人连续语音识别技术与口语对话模型实现了智能熊猫系统的人机知识问答。
It is applicable to any small vocabulary hybrid speech recognition system that combines hidden Markov model (HMM) with multi-layer perceptron (MLP).
研究适用于隐马尔可夫模型(HMM)结合多层感知器(mlp)的小词汇量混合语音识别系统的一种简化神经网络结构。
Since the widely used Hidden Markov model (HMM) in speech recognition is first order Markov model, it can not fully model the temporal dependence of speech signal.
由于在语音识别中被广泛应用的隐马尔可夫模型(HMM)是一重马尔可夫模型，它不能充分地描述语音信号的时间相依性。
Traditional speech recognition system has an intrinsic defect that, commonly only use the acoustic model of speech and unable to use non-acoustic knowledge of language to recognize speech.
仅仅依靠语音信号的声学模型来进行语音识别，存在着不能利用语言的非声学知识的固有缺陷。
A discrete hidden Markov model based on the multiple vector quantization codebooks is used here for speaker-dependent discrete speech recognition in Noisy Environments.
本文介绍多码本离散隐马尔可夫模型用于含噪声语音识别的研究成果。
But if this model is applied in speech recognition directly, it would produce the problems of rule disaster and network ratiocination invalidation.
但直接将该模型用于语音识别，将会使网络产生规则灾和网络推理失效等问题。
Parametric Stochastic Trajectory Model: in a speaker recognition system, it's often encountered that the speech data isn't enough for training.
参数化随机轨线模型：在说话人识别系统中，经常存在训练语料不足的问题。
A whole continuous speech recognition system includes four parts: feature extraction, acoustic model, language models and search algorithms, and the thesis is carried out according to them.
完整的连续语音识别系统主要包括四个部分：特征提取，声学模型，语言模型和搜索算法，本文就是根据这四个部分展开的。
The results show that the speech recognition algorithm has high recognition rate, can reduce or eliminate noise caused by the training model and the mismatch between the speech test.
结果表明：所设计的语音识别算法有很高的识别率，能减小或者消除噪声所带来的训练模型和测试语音之间的失配。
To address the problem above, the speech recognition system has been built on the basis of HTK as well as hidden markov model theory.
针对上述问题，结合隐马尔可夫模型原理，在HTK语音处理工具箱的基础上构建了中英文特定词语音识别系统。
This paper proposes two methods for speech recognition under the additive noise environment, namely dynamic adaptation multi model spectral subtraction and multi model spectral addition.
针对语音识别中的加性噪声进行研究，提出了动态自适应多模板谱减法和多模板谱加训练补偿法。
Acoustic model and speech recognition theory is the basis for building speech recognition systems.
语音的声学模型和识别理论是构建语音识别系统的基础。
The recognition technique used for the recognition of the coded speech signals is the Hidden Markov Model technique.
编码的识别技术用于识别语音信号是隐马尔科夫模型的技术。
New model improves the speech recognition rate.
新模型使语音识别率得到了改善。
Finally, it design a small-vocabulary continuous speech recognition system based on the mixed model, through experimental verify the validity and serviceability of the mixed model.
最后，完成了基于本文混合模型的非特定人小词汇量连续语音识别系统的设计和实现，通过大量试验验证了此混合模型的有效性和适用性。
The mapped algorithm, called the Mahalanobis distance, handles about 50% of its computational load in the overall speech recognition algorithm using a continuous hidden Markov model(CHMM).
将基于连续隐含Markov模型语音识别算法中占系统总运算量的50%以上的Mahalanobis距离 计算，映射为硬件实现的 模块。
In Mandarin speech recognition, this model shows a better performance and requires less memory space than the word based trigram model.
在汉语普通话连续音识别中，这个词义模型的性能优于基于词的三元文法模型，并且需要较小的存储空间。
Secondly, the hidden Markov model which is the most popular speech recognition technology has been studied in the way of speaker indendent.
其次，在非特定人语音识别技术方面，文章研究了现行最流行基于隐马尔可夫模型的非特定人语音识别技术。
A novel model-based speaker adaptation algorithm, support speaker weighting(SSW), was proposed for rapid speaker adaptation in speech recognition systems.
针对特征语音说话人自适应算法的缺陷，提出了基于结构化特征语音模型的区别性说话人自适应方法。
A novel model-based speaker adaptation algorithm, support speaker weighting(SSW), was proposed for rapid speaker adaptation in speech recognition systems.
针对特征语音说话人自适应算法的缺陷，提出了基于结构化特征语音模型的区别性说话人自适应方法。
应用推荐