基于多点模型的3D人脸姿态估计方法

改进传统的活动形状模型法,准确地提取人脸特征点后,针对人脸形状特性,使用人脸的多个特征点作为人脸模型,通过最小二乘法优化求解,精确估计3D人脸空间姿态。实验结果表明,新方法不仅可以获得稳定的姿态解,而且与同类方法比较具有良好的姿态估计精确度。     

基于H.264编码模式的快速判决算法

H.264在运动估计中引入了多种不同的编码模式,使得模式选择过程异常复杂。针对这一问题,本文提出了一个新概念——编码模式参考集,并给出了一种基于区域预测的编码模式参考集选择方法。该方法可以大大减少候选参考模式的数目,简化模式选择过程。并根据所选模式的特点给出了一种基于模式选择的自适应搜索策略。实验结果表明,本文提出的新算法在重建图像质量没有明显下降的前提下,编码速度有了显著提高。     

基于一种改进NMF算法的人脸年龄估计方法

基于人脸图像的年龄自动估计是人脸识别领域的一个重要研究方向,同时也是一个难点。对此,提出了一种改进的NMF算法来实现人脸年龄估计,该算法首先对NMF分解的基图像进行判别分析,保留最具判别力的基图像来构造子空间,然后将整体训练集图像向得到的子空间进行投影,并用RBF(radial basis function)神经网络进行训练和测试,提取包含在大多数人脸图像上的年龄信息来进行年龄估计,实验结果表明,该算法获得了较好的测试结果。     

一种V型阵列的二维波达方向估计的新方法

波达方向估计是阵列信号处理的一个重要研究方向,在雷达、通信、声纳、地震勘测等领域都有着广泛的应用前景.它已成为阵列无源探测和智能天线中的关键技术.针对二维信号,本文研究一种基于V型阵二维波达方向估计的新算法.该算法根据阵列结构的特点形成多个需要的相关矩阵,构造一个特殊大矩阵并经特征分解获得信号子空间的估计,最后利用2D-ESPRIT方法实现二维角度估计,可以解决β角兼并信号的波达方向估计问题,无需谱峰搜索且信号参数自动配对.最后用计算机仿真验证了该算法的有效性.     

梯度估计在眼睛特征提取中的应用

眼睛特征的提取在表情分析、人脸合成和人脸识别等应用中起着非常重要的作用.目前广泛使用的眼睛特征提取算法是基于模板匹配的算法和基于特征点的算法.然而基于模板匹配的算法通常需要进行多个参数的选择,而且匹配过程非常耗时;特征点的算法则存在准确度较低的问题.针对以上问题提出了一种基于梯度信息提取虹膜中心、虹膜外轮廓和眼睛角点等眼睛特征的算法,算法用改进霍夫变换提取虹膜中心和外轮廓.在获得虹膜中心点后,参照中心位置选择眼睛角点候选区域并对其梯度进行分析.依据角点附近梯度分布与其它区域梯度分布不同的性质定位角点.仿真结果表明提出的算法能有效的提取眼睛特征,在测试数据上得到了比其它算法更高的准确率和更快的速度.     

改进周期图法功率谱估计中的窗函数仿真分析

周期图法是功率谱估计的一种基本方法,但该方法不满足一致性估计条件,谱估计的分辨力和方差都很难满足实际应用需要.因此采用其改进方法(即Welch法)估计信号功率谱,并结合该方法对四种典型窗函数的谱估计性能进行了具体的仿真,就其中的频率分辨率、采样信号的数据长度与窗函数的特性等因素之间的关系,进行了详细的讨论,重点分析了窗函数对谱估计的影响,指出了它们的优缺点,得出了不同数据类型与估计方式下的功率谱标准方差的均值.最后,根据实验结果并结合窗函数的评价指标,指出了选取合适的窗函数应注意的一般原则,具有一定的实际指导意义.     

网络蠕虫感染率的抗差估计法

网络蠕虫对信息安全构成了威胁,检测和防范蠕虫成为网络安全的研究课题。本文提出了一种网络蠕虫感染率的估法模型,然后根据权函数和残差区间给出了抗差等价权矩阵,分析了误差影响及验后精度估计,最后进行了仿真实验。实验结果表明,这种方法具有良好的抗粗差能力,可靠靠且收敛速度快。     

An auxiliary theorem for stability analysis in the presence of interval-valued parameters

A theorem which helps to determine the boundary of the image of an interval inR ⁿ under a differentiable mapping,F:R ⁿ C, is expounded. Examples illustrating the application of the theorem are given.     

Infinitesimal perturbation analysis of a queueing system with bursty traffic

We study aG/G/1 queueing system with a bursty arrival process. Based on a general model for such a bursty process, we derive infinitesimal perturbation analysis (IPA) derivative estimators of the mean system time with respect to various parameters of interest. The cases of both complete and partial state information are considered. To ensure unbiasedness and strong consistency of the estimators, different sample path representations are developed such that sample functions are continuous with respect to the particular parameter of interest. Some of these representations are applicable to a wider class of gradient estimation problems where sample path discontinuities arise. Simulation results are included to compare the convergence rates and variance properties of the different IPA estimators developed.     

Analyses of (AD)-sorption for the estimation of pore-network dimensions and structure

The interpretation of pore dimensions based on physical ad-desorption analyses is central to the characterization of pore network structure. Several approaches have been proposed and are commonly employed in the analysis of physical adsorption and/or desorption to deduce the dimensions of the porous network. These approaches assume either theoretical (e.g., BET, the Halsey equation as interpreted by Pierce et al., or the more recent analyses of microporosity) or standard isotherms as model(s) for the sequential calculations required in estimating the pore network dimensions. Subsequent representation of the pore dimensions and the relationship between these distributions in dimension and other experimental parameters (such as catalytic activity, adsorptivity or transport); thus, depend explicitly on the model employed in the analyses. Each instrument currently available for the measurement of porous solid structure by sorption employs the same specific models for the relationship between the volume ad-desorbed and the dimensions of the porous network that is being characterized.This paper analyzes the interpretation of porous dimensions based on the sequential calculations required in the analyses. A new approach is proposed which is based on a modification to current practices reflecting Halsey's original theory for the thickness of the adsorbed layer (as a function of P/P ₀). Further, the calculations of the incremental changes in the exposed surface area are discussed as they relate to pore network structure. A method is proposed to infer the differences in pore shape. Sorption data are analyzed by these new approaches, and these analyses will be compared with those approaches currently employed. Analyses based on these modified approaches provide a dramatically more consistent interpretation of the sorption data and the corresponding pore network structures.