一种基于音质评估的语音伪装通信新方法

研究一种语音伪装新方法,将保密语音替换成另外一种语音以便于在公网上安全传输。保密语音经2.4kb/低速率压缩编码后,依据设计的客观音质评估模型,将参数隐蔽地嵌入普通语音;检测时利用逆过程恢复保密语音。实验表明算法的信噪比较传统算法提高约3.5dB,并在受到压缩、滤波等攻击时具有较高的鲁棒性。     

改进的基于信号子空间的多通道语音增强算法

通过同时对角化麦克风阵列接收信号中语音信号和噪声信号的全局协方差矩阵,本文改进了一种基于信号子空间分解的多通道语音增强算法.该算法不依赖任何信号模型且无需对噪声信号的统计特性进行任何先验假定,它弥补了原始算法只限于白噪声背景下语音增强的不足,实现了色噪声背景下语音信号的最优估计.仿真结果表明本文算法在主观和客观测试中都具有良好的语音增强效果.     

一种基于小波变换的独立语音量分离方法

提出了一种基于小波变换的独立语音量分离方法,以改进卷积叠加语音分离方法.该方法在小波域进行独立元分析,把高阶去相关矩阵分解为小波子空间中一系列相应的低阶去相关矩阵,并在保证语音短时平稳性的基础上通过低阶处理跟踪时变的室内(存在室内回响)传输函数.实验结果表明,提出的方法改善了传统独立元分析在室内回响环境中独立语音量分离的效果.     

一种基于广义奇异值分解的语音增强算法

广义奇异值分解的单通道语音增强算法在加性噪声为白噪声的情况下,效果比较理想.加性噪声为有色噪声的情况下,通常用一种基于熵奇异值分解(QSVD)的方法来处理.对QSVD算法进行衍生,首先提出了一种基于广义奇异值分解的子空间语音增强算法(GSVD).为了处理低信噪比时残留的音乐噪声,结合入耳的听觉掩蔽效应,进一步提出了一种基于感官抑制的GSVD(PCGSVD).试验结果显示,PCGSVD算法能够明显地提高语音质量、可懂度和识别率,特别是在加性噪声是有色噪声的情况下实验结果明显优于其他的语音增强算法.     

Rationale for a Modified Duane Model

The Duane model for reliability growth involves a rate function which is an inverse power law and has an ``infinite' value at t = 0. The model is usually motivated entirely empirically. Here a probabilistic rationale is proposed via a reliability growth model involving the removal of design faults. This rationale results in a modified power law rate, finite at the origin. A wider class of rate functions should be investigated for NHPP models of reliability growth.     

噪声自适应的多数据流复合子带语音识别方法

首先针对现有丢失数据语音识别技术中的边缘化(marginalisation)技术在特征运用上的局限,提出了一种倒谱特征分量的可靠性估计方法,将边缘化技术推广到常用的倒谱语音识别系统中; 然后利用基于全带和子带倒谱特征的边缘化识别器在不同噪声中的互补性能,提出了一种噪声自适应的多数据流复合子带语音识别方法。实验结果表明,所提识别方法可以自适应地选出全带和子带数据流中受噪声影响较小者并以之为主要依据进行识别,有效地提高了识别系统在多变噪声环境中的鲁棒性。     

Metrics for Power Spectra: An Axiomatic Approach

We present an axiomatic framework for seeking distances between power spectral density functions. The axioms require that the sought metric respects the effects of additive and multiplicative noise in reducing our ability to discriminate spectra, as well as they require continuity of statistical quantities with respect to perturbations measured in the metric. We then present a particular metric which abides by these requirements. The metric is based on the Monge-Kantorovich transportation problem and is contrasted with an earlier Riemannian metric based on the minimum-variance prediction geometry of the underlying time-series. It is also being compared with the more traditional Itakura-Saito distance measure, as well as the aforementioned prediction metric, on two representative examples.      

谱相减与波束形成相结合的语音增强算法

根据阵列信号语音增强的思想,提出一种基于频城处理的谱相减与波束形成相结合的语音增强结构。结构为多路信号输入,每路含噪信号在谱相减后,增加了波束形成结构,不仅有效地消除了背景噪声,也抑制了谱相减后的音乐噪声。并使用该算法对实际环境中采集到的含噪语音信号进行了仿真,结果显示经过该系统处理后的增强语音的信噪比有了较大的提高,主观试听效果也很好。     

一种快速稳健的语音／音乐分类方法

介绍了一种简单而有效的区分语音/音乐的方法.该方法仅提取唯一的区别性特征,即改进型低能量帧率(MLER),应用贝叶斯最大后验分类器判断每1段的音频类别,再根据音频流的相邻段有很强的相关性这一特点,采用一种新颖的基于相关性的后验决策方法对分类器的分类结果进行进一步改善.实验结果表明,该方法算法简单,分类效果好.     

A Subspace Projection Approach for Analysis of Speech Under Stressed Condition

In this paper, a novel subspace projection approach is proposed for analysis of speech signal under stressed condition. The subspace projection method is based on the assumption of orthogonality between speech subspace and stress subspace. Speech and stress subspaces contain speech and stress information, respectively. The projection of stressed speech vectors onto the speech subspace will separate speech-specific information. In this work, the speech subspace consists of neutral speech vectors. Speech and stress recognition techniques are used to verify the orthogonal relation between speech and stress subspaces. The evaluation database consists of 119 word vocabulary under neutral, angry, sad and Lombard conditions. Hidden Markov models for speech and stress recognition are used with mel-frequency cepstral coefficient features for evaluation of estimated speech and stress information.     

一种语音特征参数子分量分析与有效性评价的新方法

语音信号中包含语义和说话人个性两大特征,其有效提取和强化对语音识别和说话人识别有着非常重要的意义。本文提出了一种语音特征参数中语义和个性特征子分量分析与有效性评价的4S方法,对语义和个性特征的成份比例进行分析,并通过量化指标评判特征参数对语音识别和说话人识别的有效性。运用4S分析方法对目前常用的特征参数LPC, LPCC和MFCC的子分量分析与有效性评价结果表明,所有的特征参数都更多地包含了语义特征信息,语义特征和说话人个性特征的成份比例因子LIR分别为1．30、1．44和1．61,并且,三种参数对语音识别和说话人识别的有效性均呈现出依次提高的特性。     

On a Speech Multiple System Implementation for Speech Synthesis

This paper proposes a voice synthesizer to convert a single speech to multiple speeches. Pitch is an important voice characteristic of speech parameter and it is produced by the periodic vibration of the vocal-cords; the parameter most sensitive for human’s auditory sense. So if you change the pitch of the voice with several different scaling rates, you can produce several different voices at the same time with a voice. The Multiple-Speech Synthesizer will be used in diverse practical applications. The cheer synthesizer that makes group voice from a single voice would be such an example. You could also use the system for a troll toy, birthday song tracks, effect sounds in movies and plays, protection systems of houses, etc. Furthermore, the synthesizer could be used to imitate the voice of famous entertainers or cartoon characters, such as the Mask-man.

An Efficient Approach for Enhancing Contrast Level and Segmenting Satellite Images: HNN and FCM Approach

Satellite image segmentation has gotten bunches of consideration of late because of the accessibility of commented on high-goals image informational indexes caught by the last age of satellites. The issue of fragmenting a satellite image can be characterized as ordering (or marking) every pixel of the image as indicated by various classes, for example, structures, streets, water, etc. In this paper centered to build up a satellite image segmenting process by utilizing distinctive optimization methods. The work is prepared dependent on three stages that are RGB change, preprocessing, and division. At first the database images are assembled from the database at that point select the blue band images by performing RGB change. To improve the differentiation and furthermore decreasing the commotion of these chose blue band images, Hopfield neural network (HNN) is utilized. After image upgrade, the images are fragmented dependent on fuzzy C means (FCM) clustering method. The images are clustered and segmented in the way of optimizing the centroid in FCM utilizing oppositional crow search algorithm. The exhibition of the proposed framework is investigated dependent on the presentation measurements, for example, affectability, particularity and accuracy. From the outcomes, the proposed strategy diminished the computational time by expanding the accuracy of 98.3% with HNN system.

     

一种适于非特定人语音识别的并行隐马尔可夫模型

为了适合非特定人语音识别,提出了一种由多条并行马尔可夫链组成的并行HMM(Parallel Hidden Markov Model,PHMM),从而融合了基于分类的语音识别中为各个类别建立的模板,提高了识别性能,各条链之间允许有交叉,使得融合的多模板之间存在状态共享,同时PHMM可以在训练过程中自动完成聚类,且测试语音的输出结果来自所有类别,无需聚类分析和类别判断,这些都减少了存储量和计算量,汉语非特定人孤立数字的识别实验表明,PHMM较之传统CHMM使识别性能及噪声鲁棒性都得到了改善。     

一种甚低速率语音编码的抗误码算法

为满足在高误码率的窄带信道上进行语音通信的需求，本文研究了一种适用于甚低速率语音通信的抗误码参数估值算法。在一定的解码状态下，声码器通过计算最小均方误差（MMSE）估计的方法估计最优参数，充分降低信道误码对重建语音质量的影响。对于解码状态参数，通过计算最大后验转移概率的方法作最佳估计，并给出了一种简化的计算方法。这种抗误码算法复杂度低。计算机仿真结果表明，在不同误码率下该算法恢复出的语音的平均谱失真（ASD）低于帧删除方法的谱失真最大可达0.4dB。     

路由选择的智能方法

本文利用模糊理论来对通信网中的路由进行动态选择。文中首先讨论了路由选择所要考虑的一些因素,然后利用模糊综合评判理论评估出路由选择的量化综合指标,再通过模糊图中的最强路法选择出两点间的最佳路由。最后把该法同已有方法进行仿真比较,发现该法比较的方法能较大地提高接通率,且性能未受损害。     

Wise-ShopFloor: An Integrated Approach for Web-Based Collaborative Manufacturing

This paper presents an integrated approach for Web-based collaborative manufacturing, including distributed process planning, dynamic scheduling, real-time monitoring, and remote control. It is enabled by a Web-based integrated sensor-driven e-ShopFloor (Wise-ShopFloor) framework targeting distributed yet collaborative manufacturing environments. Utilizing the latest Java technologies (Java 3D and Java Servlet) for system implementation, this approach allows users to plan and control distant shop floor operations based on runtime information from the shop floor. The objective of this research is to develop methodology and algorithms for Web-based collaborative planning and control, supported by real-time monitoring for dynamic scheduling. Details on the principle of the Wise-ShopFloor framework, system architecture, and a proof-of-concept prototype are reported in this paper. An example of distributed process planning for remote machining is chosen as a case study to demonstrate the effectiveness of this approach toward Web-based collaborative manufacturing.     

基于特征空间逼近的传声器阵列语音信号增强

对现有的基于自动波束形成的传声器阵列语音信号增强算法提出了改进,将各传声器采集到的信号利用ABF(自适应波束形成)进行延时补偿并求和,消除信号中弱相干和不相干噪声;再利用特征空间逼近的方法进一步去除残留的噪声。将一种定阶方法应用到基于特征空间分解的语音信号增强中,利用其“最大稳定”原理,使得有效信号模型的阶数尽可能不受原始信号信噪比的影响,消除了传统定阶过程中的随意性和不稳定性。仿真结果表明:把自适应波束形成技术和特征空间逼近的方法结合起来,能够取得良好的去噪效果。     

基于最小统计及谱减法的语音增强

提出了一个基于最小统计及谱减法的语音增强方法。不采用端点检测,在语音帧内及噪声帧内都进行噪声更新。为了实时应用的需求,提出了一种加快最小值更新的方法,实验表明,该方法能有效削弱变电平噪声。