基于SURF与Hough森林的人脸检测研究

为实现复杂场景中的人脸检测与定位,提出了一种基于快速鲁棒特征SURF与Hough森林的人脸检测算法。采用SURF局部特征构建Hough森林分类器,每个叶子节点存储类别信息与特征点到达目标中心的偏移量,在图像局部外观与Hough投票之间建立映射关系,生成有监督的判别式的码本,获得可靠的概率Hough投票,以此预测目标中心位置,提高了检测精度。与此同时,采用SURF局部特征提取图像兴趣点有助于减小计算量、加快检测速度。实验证明了所提算法的有效性。     

An improved t‐out‐of‐n e‐lottery protocol

In 2009, Lee and Chang proposed an electronic t‐out‐of‐n lottery protocol over the Internet based on the Chinese remainder theorem and blind signature. However, a security flaw exists in Lee–Chang's protocol that the lottery agent and the malicious purchaser can collude to control the winning result, which is unfair for the honest participants. On the basis of a verifiable random number generated by using the Lagrange interpolation formula over a finite field, an improved t‐out‐of‐n e‐lottery protocol is proposed, which guarantees that each participant can generate the winning result equally. The improved protocol achieves public verifiability and fairness without a trusted third party and a delaying function, which makes it more robust and more efficient. Copyright © 2013 John Wiley & Sons, Ltd.     

Compressed sensing reconstruction of sparse spectrum based on digital micro-mirror device platform

A new method which employs compressive sensing (CS) to reconstruct the sparse spectrum is designed and experimentally demonstrated. On the basis of CS theory, the simulation results indicate that the probability of reconstruction is high when the step of the sparsity adaptive matching pursuit algorithm is confirmed as 1. Contrastive analysis for four kinds of commonly used measurement matrices:part Hadamard, Bernoulli, Toeplitz and Circular matrix, has been conducted. The results illustrate that the part Hadamard matrix has better performance of reconstruction than the other matrices. The experimental system of the spectral compression reconstruction is mainly based on the digital micro-mirror device (DMD). The experimental results prove that CS can reconstruct sparse spectrum well under the condition of 50% sampling rate. The system error 0.078 1 is obtained, which is defined by the average value of the 2-norm. Furthermore, the proposed method shows a dominant ability to discard redundancy.     

运用高斯混合模型识别动物声音情绪

针对动物的情绪识别问题,提出高斯混合模型在动物声音情绪识别上的应用方法。利用语音信号处理与机器学习技术,提取动物声音信号的过零率、共振峰、梅尔 频率倒谱系数3种描述动物情绪的特征参数。采用高斯混合模型对采集到的动物声音信号训练样本进行聚类分析,计算测试样本后验概率,实现动物情绪的自动识别。通过分析特征参数的权重系数组合、高斯混合数目对识别率的影响来择选最优参数。实验结果表明,经参数优化后的高斯混合模型可将动物声音情绪的识别率由84.25%提高至96.67%。     

Microstructure and optoelectronic properties of gallium- titanium-zinc oxide thin films deposited by magnetron sputtering

Gallium-titanium-zinc oxide(GTZO) transparent conducting oxide(TCO) thin films were deposited on glass substrates by radio frequency magnetron sputtering. The dependences of the microstructure and optoelectronic properties of GTZO thin films on Ar gas pressure were observed. The X-ray diffraction(XRD) and scanning electron microscopy(SEM) results show that all the deposited films are polycrystalline with a hexagonal structure and have a preferred orientation along the c-axis perpendicular to the substrate. With the increment of Ar gas pressure, the microstructure and optoelectronic properties of GTZO thin films will be changed. When Ar gas pressure is 0.4 Pa, the deposited films possess the best crystal quality and optoelectronic properties.     

面向管理实体对象的海量工程项目数据模型及实现

针对建立统一数据模型的PIP数据中心的需求,提出一种面向管理实体对象的数据基本模型,把工程项目数据按项目全生命周期过程分解为若干个基于工作流的可管理的实体对象,给出了项目工程数据概念层模型,利用该模型实现项目工程数据的自然表达和记录。进一步,基于文档型数据库MongoDB技术提出工程项目数据的组织方法,并设计了存储服务集群架构方案。实验表明,该存储方案具有高效的数据写入和查询性能,具有服务高可用性和存储能力的可扩展性等特点。     

Effect of working pressure on the structural, optical and electrical properties of titanium-gallium co-doped zinc oxide thin films

The transparent conducting titanium-gallium co-doped zinc oxide (TGZO) thin films were grown on glass substrates by radio-frequency magnetron sputtering technique. The effects of working pressure on the structural, optical and electrical properties of the films were investigated. The results show that the deposited films are polycrystalline with a hexagonal wurtzite structure and highly textured along the c-axis perpendicular to the substrate. The TGZO film prepared at the working pressure of 0.4 Pa exhibits the best crystallinity, the maximal grain size, the highest transmittance, the lowest resistivity and the highest figure of merit. The optical constants of the films were calculated using the method of optical spectrum fitting. The dispersion behavior of the films was studied by the single-electronic oscillator dispersion model. The oscillator parameters and optical bandgaps were determined. The results demonstrate that the microstructure and optoelectrical properties of the TGZO films are dependent on the working pressure.     

Performance study of hybrid spectrum access scheme in millimeter wave cellular network

In millimeter wave cellular networks,the spectrum efficiency of multi-operator dual frequency hybrid spectrum access scheme under open access mode was studied.Firstly,Poisson point process was used to model the distribution of base stations of each operator,the base station and carrier frequency were selected by users based on maximum received power criterion and high frequency carrier signal to interference plus noise ratio (SINR) threshold access criterion respectively.Secondly,the rate coverage was used as performance criteria to study the spectrum efficiency,and the theoretical expression of rate coverage was derived based on stochastic geometry theory according to channel model,path loss model and antenna model.Finally,the influence of user density,base station density and antenna gain on rate coverage was analyzed by simulation.The experimental results show that the multi-operator dual frequency hybrid spectrum access scheme has better spectrum utilization efficiency.     

基于MPSoC并行调度的矩阵乘法加速算法研究

矩阵乘法是数值分析以及图形图像处理算法的基础,通用的矩阵乘法加速器设计一直是嵌入式系统设计的研究热点。但矩阵乘法由于计算复杂度高,处理效率低,常常成为嵌入式系统运算速度的瓶颈。为了在嵌入式领域更好地使用矩阵乘法,提出了基于MPSoC(MultiProcessor System-on-Chip)的软硬件协同加速的架构。在MPSoC的架构下,一方面,设计了面向硬件约束的矩阵分块方法,从而实现了通用的矩阵乘法加速器系统;另一方面,通过利用MPSoC下的多核架构,提出了相应的任务划分和负载平衡调度算法,提高了并行效率和整体系统加速比。实验结果表明,所提架构及算法实现了通用的矩阵乘法计算,并且通过软硬件协同设计实现的多核并行调度算法与传统单核设计相比在计算效率方面得到了显著的提高。     

基于因子图的分布式变分稀疏贝叶斯压缩感知简

提出了一种基于因子图的分布式变分稀疏贝叶斯压缩感知算法。该算法利用因子图和变分方法将全局感知问题分解为简单的局部问题,通过认知用户邻居间的置信传播实现“软融合”,使每个认知用户能够获得全局最优估计。且充分利用邻居间传递的信息所具有的时间和空间二维相关性,提高认知用户在低信噪比下的感知性能。同时,算法在迭代过程中自适应地删除不收敛的超参数及对应的基函数,降低通信负载。实验结果表明：该方法在低采样率和低信噪比下有较好的感知性能。