基于多重分形谱的木材高光谱图像纹理分类算法

为了过滤木材高光谱图像中大量的冗余信息,提升应用图像纹理进行分类的准确率,本文采用基于多重分形理论的木材高光谱图像分类算法。首先利用不同的特征选择算法选取最具代表性的10个波段;随后根据不同的函数密度图像对所选取波段的图像求解其多重分形曲线,将选择出的多个波段所对应的多重分形曲线取平均,得到表示样本纹理特征的多重分形曲线;最后使用支持向量机和BP神经网络分类器对多重分形曲线进行分类。实验表明,相对熵(K-L散度)要好于自适应波段选择(ABS)提取的波段,多重分形算法提取的高光谱图像纹理特征要好于灰度共生矩阵,支持向量机算法的分类准确率和速度要优于BP神经网络,融合K-L散度、多重分形和支持向量机算法能够有效地提高木材高光谱图像的识别准确率,最高识别准确率达到了97.91%。     

基于小波分解和支持向量机的准正面人脸识别方法

基于小波分解提取人脸特征技术和多分类支持向量机模型，提出了一种新的准正面人脸识别算法。小波分解提取人脸特征具有对表情变化不敏感的特点；支持向量机作为分类器被认为具有很高的推广(generalization)性能，无需先验知识。在所提出的算法中，首先对训练图像进行预处理，然后使用小波分解方法对人脸图像进行特征提取，用所提取的人脸特征向量训练多分类支持向量机模型，最后用训练好的支持向量机进行人脸识别。利用ORL人脸图像库对该算法的实验测试结果，以及与其它人脸识别方法的比较结果表明了该算法在识别性能方面的优越性。     

最小二乘支持向量机的一种非均衡数据分类算法

为了提高支持向量机的非平衡数据分类能力,分析了最小二乘支持向量机的本质特征,提出了一种基于中心距离比的非平衡数据分类算法,同时通过修剪边界样本,解决了最小二乘支持向量机缺失稀疏性的问题.在UCI标准数据集上进行的试验表明:该算法能够有效地提高支持向量机对非均衡分布数据的正确性,且该算法在不影响训练精度的前提下,可以得到稀疏解,算法的训练速度也有了一定的提高.     

基于次特征值误差补偿和非对称分布的马氏距离改进算法

本文提出了一种有限样本集上基于次特征值误差补偿和优势主向量上非对称分布的马氏距离改进算法.通过改进的马氏距离,有限样本导致的次特征值误差得到补偿,样本特征矢量在变换空间的各优势主向量上的投影分布得到更精确的刻画,因此可以有效地计算最近邻参考矢量.在UCI手写体数字字符数据库上的识别实验结果表明,该改进算法对于提高识别性能是有效的.     

基于双谱奇异值和最优参数支持向量机的目标识别算法

提出了基于双谱奇异值和最优参数支持向量机的目标识别算法。首先从目标距离像的双谱中提取双谱奇异值特征;然后通过实验的方法确定支持向量机参数取值的范围,利用遗传算法获取支持向量机的最优参数;最后应用最优参数支持向量机对3种地面坦克目标进行识别。仿真实验表明,即使在较低的信噪比下该算法仍然可以取得比较高的识别率。     

小波变换和支持向量机相融合的 ECG 身份识别算法

心电图信号(ECC)受到设备、检测者心情等影响,含有噪声且特征具有高维性,为了提高了身份识别的正确率,提前出一种小波变换和支持向量机相融合的 ECG 身份识别算法.首先小波变换对采集的心电图信号进行预处理,消除各种噪声;然后采用相关分析对 ECG 特征进行降维处理,消除冗余特征;最后将特征向量输入到支持向量机进行训练,建立基于 ECG 的身份识别模型.仿真结果表明,该方法提高了身份识别正确率和速度,有一种有效的身份识别算法.     

管道缺陷漏磁信号分类与识别方法的研究

为了提高管道漏磁检测缺陷识别的准确性和识别效率,引入支持向量机作为缺陷分类与识别的平台.利用ANSYS软件对具有不同长度、宽度和深度的多组缺陷模型进行缺陷漏磁场仿真,将仿真得到的漏磁场磁通密度数据保存在文本文件中,封装成符合LIBSVM支持向量机工具格式的训练数据,对支持向量机训练.使用训练后得到的支持向量机模型进行缺陷分类.实验结果表明,支持向量机分类性能良好,应用在漏磁检测缺陷分类中具有可行性.     

基于GWO-SVM的红外热成像低零值绝缘子识别

为了准确识别电网中的低零值绝缘子,提高劣化绝缘子诊断的准确率,提出了一种使用灰狼算法优化（grey wolf optimizer, GWO）与二进制支持向量机（support vector machine, SVM）分类器相结合的模型,对红外图像中的低零值绝缘子进行自动检测。首先对绝缘子红外图像进行增强,利用Ostu算法对红外图像进行分割,并对得到的二值图像进行倾斜角度矫正和切割,提取绝缘子串的有效区域,然后将图像特征用于向量机的分类识别。实验结果表明,灰狼算法优化支持向量机比常用的网格搜索算法（grid search, GS）、粒子群优化算法（particle swarm optimization, PSO）等得到的分类模型能更准确、有效地对低零值绝缘子进行识别,且准确率更高。     

用改进核函数提高SVM的雷达目标识别率

对支持向量机中的高斯核进行了改进,利用改进的高斯核构造了一维高分辨率距离像的雷达目标识别算法,并将幂变换引入预处理过程.该技术提高了识别率,减少了识别时间;同时对所完成的目标识别算法的性能进行了评估,从方位角大小、信噪比和训练数据大小三个方面验证了该算法的稳健性.     

基于支持向量机的目标图像识别技术

介绍了目标图像的前期处理、目标图像的特征提取和基于支持向量机的目标图像识别方法。为了验证方法的正确性．采用了一批太空目标图像对其算法进行了检验。数据实例表明，支持向量机识别的正确率比其它两种神经网络方法的正确率要高很多。     

张量正交局部敏感判别式分析及其在人脸识别中的应用

针对人脸识别的特征提取问题,本文提出了一种张量正交局部敏感判别分析(Tensor-based Orthogonal Locality Sensitive Discriminant Analysis, Tensor-OLSDA)的人脸识别算法。张量正交局部敏感判别分析在保持了流形的局部几何结构的同时加强了全局判别结构,并克服了局部敏感判别分析算法中非正交性带来的度量失真和维数估计困难等问题,从而增强了数据的可分性,提高了识别效果。张量正交局部敏感判别分析首先将人脸数据表示成高阶张量形式,在进行特征提取时将高阶张量数据沿不同阶展开,再利用特征根之间的正交性约束条件,求解正交局部敏感判别式分析特征子空间,最后将高阶人脸数据投影于这个特征子空间,进行识别。在AT&T和YaleB人脸库上的实验结果表明,Tensor-OLSDA具有良好的分类性能,能获得较为理想的识别结果。      

Robust$H_infty$Power Control for CDMA Cellular Communication Systems

Power control is an important factor for direct-sequence code division multiple access (DS-CDMA) cellular radio systems to achieve higher communication link quality and better system capacity. In order to track the desired signal-to-interference-plus-noise ratio (SINR) under round-trip delay, multiple access interference (MAI), channel fading, and noise, a time delay-based state-space model is developed for representing the tracking error dynamics and a state feedback controller is introduced for SINR tracking control. Then the power tracking problem can be regarded as a control problem. In this paper, a robust$H_infty$power tracking control design is proposed to achieve a robust optimal SINR tracking from the minimization of the worst-case effect point of view. This robust optimal power tracking design problem can be transformed to solving the eigenvalue problem (EVP) under some linear matrix inequality (LMI) constraints. The LMI Matlab toolbox can be used to efficiently solve the EVP via convex optimization to achieve a robust optimal SINR tracking design. Under the proposed distributed framework, the information of channel gain is not needed.     

加性噪声条件下鲁棒说话人确认

基于非负矩阵分解的语音去噪,在提高语音信号信噪比的同时,也会引起语音失真,从而导致噪声环境下说话人确认系统性能下降.本文提出基于分区约束非负矩阵分解的语音去噪方法（Nonnegative Matrix Factorization with Partial Constrains,PCNMF）,目的是在未知和非平稳噪声条件下提高话人确认系统的鲁棒性.PCNMF在满足分区约束条件的基础上分别构建语音字典和噪声字典.考虑到传统语音训练产生的语音字典往往含有一定的噪声成分,PCNMF通过数学模型产生基音及泛音频谱,在此基础上利用该频谱模仿人声的共振峰结构来合成字典,从而保证语音字典纯净性.另一方面,为了克服传统噪声字典构建方法带来的部分噪声信息丢失问题,PCNMF对在线分离出的噪声样本进行分帧和短时傅里叶变换,然后以帧为单位线性组合生成噪声字典.性能评估实验引入了多种噪声类型,实验结果表明PCNMF可有效提高说话人确认系统的鲁棒性,特别是在未知和非平稳噪声条件下其等错率相比基线系统（Multi-Condition）平均降低了5.2%.     

Real-Time Implementation of Time Domain Harmonic Scaling of Speech for Rate Modification and Coding

Time domain harmonic scaling (TDHS) has been realized in real time on the Bell Laboratories digital signal processing (DSP) integrated circuit. It is an algorithm that can expand or compress the bandwidth and sampling rate of speech by taking advantage of the pitch structure in the speech signal. As such it is useful in a variety of speech applications including speech coding, speech enhancement, and rate modification. A single DSP can perform compression and a second DSP can perform expansion. Both operations require pitch information to be supplied with the input speech. Included in the system is a real-time pitch/periodicity detector which has also been implemented on a single DSP. Its design is based on a novel modification of the autocorrelation function type pitch detector. This paper presents details of both the TDHS and pitch detector implementation and discusses their performances. In particular in this paper we discuss a 2:1 compression and expansion system that has been used as part of a 9.6 kbit/s speech coder. TDHS was previously thought to require a much larger buffer than the RAM memory available in the DSP. We show that for all the compression/expansion ratios of interest the buffer size needed is twice the maximum pitch period.     

Transformation of whispering voice to pseudo-real voice for unvoiced telephony and communication aid for voice-handicapped persons

For voice handicapped people, an easy to use voicing aid device is wanted. In mobile telephony, so-called non-speaking speech communication is an expected solution for essential privacy as well as for acoustic nuisance prevention. The study introduced here intends to cover both issues, introducing a system where the whispering (non-speaking voice or talk without vocal fold activation) signal is converted to pseudo-real voice signal, which is to be sent to, or heard by, the other party. The study also includes validation tests with multiple volunteers for its output legibility. Unlike general concept of speech regeneration being inclined to signal recognition or decomposition to text followed by electronic reading (voicing), our system converts it almost directly without recognition or decomposition steps. The processing is based on repetitive playback of short time autocorrelation, conducted by synthetic pitch pulse. A real time software pipeline process is under development.     

基于凌阳单片机的语音识别系统设计

语音识别作为一种人机交互技术,有着十分广泛的应用。介绍一种基于具有DSP（数字信号处理器）功能的凌阳单片机的语音识别系统设计,它能完成针对特定人的若干命令（特定的孤立词）的语音识别。本设计在具体实践时,以汉语言的短命令为识别对象,体现汉语音特征;在系统软件和硬件设计上强调简单实用的特点,简化系统结构,增强控制能力。文中介绍的方法克服了使用DSP芯片建立同样系统时的一些不足,如设计繁琐、价格高、控制不够灵活等。     

一种用于3G系统中复杂背景噪声环境下的话音激活检测算法

本文讨论了一种新的应用在3G自适应多速率系统中复杂背景噪声环境下的话音激活检测算法。这种算法基于谱估计理论和周期信号检测方法,应用一个IIR滤波器组把输入窄带语音信号分面九个频带,进而估计每个频带的语音信号和背景噪声的电平,结合基音和音调检测,在区分语音和移动环境中的大多数背景噪声对表现得足够健壮,最后,基于欧洲电信标准委员会建议的3G平台对这种算法进行了仿真评估,并就其健壮性从主客观两个方面进行了分析。     

Enhancement of the noise immunity of a voice-activated robotics control system based on phonetic word decoding method

The problem of protection of a voice-activated robotics control system against the acoustic noise similar to the speaker voice is investigated. A new algorithm for voice signal processing using the phonetic word decoding method and the (R+1) information element is proposed. The results of the experimental study of this algorithm are reported. It is shown that the developed algorithm reduces the false alarm ratio by and an order of magnitude or even more and retains high probability of correct voice recognition.     

Robust signal selection for the matched filter

A matched filter's performance is strongly related to the signal being detected and can be shown to be optimal when the signal is an eigenvector of the noise correlation matrix corresponding to a minimum eigenvalue. When fewer correlations are known than would be necessary to specify such an eigenvector, it is natural to choose a signal which is robust to the implied uncertainty in the noise dependency structure. This is shown to be tantamount to finding a tight upper bound on the minimum eigenvalue over all correlation matrices within the uncertainty class. Such a bound is achieved by the reduced correlation matrix of order equal to the number of available correlations, and hence the robust signal is shown to have this length. No matter how reasonable, any assumption used to extend the correlation matrix can degrade performance; a system designer should not try to use information that is not available     

异构携能通信网络中基于人工噪声辅助的鲁棒安全传输方案

异构携能通信网络中信道状态信息不准确时,为保证信息、能量传输的安全性与可靠性,该文提出一种基于人工噪声辅助的鲁棒安全传输方案。通过联合设计宏基站、微基站的下行信息波束及人工噪声矩阵,干扰窃听者,同时提升系统能量接收性能。在基站的发送功率约束、合法用户的信息接收及能量接收中断约束和窃听者的窃听信息中断约束下,以最大化系统能量接收性能为目标进行建模。针对该问题的非凸性,首先将其等效转化为一种易于处理的形式;而后进一步利用Berstein-type不等式处理其中的中断概率约束,将其转化为凸的问题进行求解。仿真结果验证了该方案的安全性和鲁棒性。