基于生物视觉标准模型特征的无参考型图像质量评价方法

鉴于生物视觉特征对于图像的良好表征能力,提出了一种基于生物视觉特征的无参考型图像质量评价方法。对生物视觉ST模型进行了研究和分析,完成了对图像的稀疏化表示;利用最小二乘支持向量机回归方法训练生物视觉特征到图像质量的映射关系,获得能够预测图像质量的回归器;通过学习的回归器完成了对图像质量的评价。基于LIVE图像库的实验结果表明,该方法对于特定失真和交叉失真的预测误差分别为2%和5%左右,并且与目前技术条件下的质量评价方法相比具有很好的精确性和单调性。     

小波不变矩的低空风切变识别

针对微下击暴流、低空急流、顺逆风以及侧风低空风切变样本图像间的形状特性关系,主要研究了小波不变矩的特征提取技术在风切变识别中的应用.首先,采用基于三次B样条的小波不变矩提取风切变图像的形状特征.然后,将提取的特征通过Fisher线性判别分析(LDA)降低维数,实现风切变有效特征的提取.最后,采用三阶近邻分类器分类识别四种低空风切变.实验结果表明,该算法与应用Hu矩和Zernike矩特征进行分类识别相比,识别结果更加稳定,且平均识别率得到了较大提高,能够有效用于风切变图像的类型识别中.     

基于多维特征的密集转发式干扰识别方法

针对密集转发式干扰与真实目标回波的高相干性,传统基于单维特征的干扰识别方法识别率低的问题,提出了基于多维特征的干扰识别方法。从分数阶傅里叶变换域提取调频参数,从自相关幅度谱和匹配滤波幅度谱分别提取矩峰度系数,从频域提取频谱半带宽方差比系数,组成特征向量,作为决策树和BP神经网络的输入,对几类密集转发式干扰进行分类识别。仿真结果表明,所提特征参数对噪声不敏感,具有较好的稳定性和分离性,在干信比大于6dB时,能够有效辨识几类密集转发式干扰。     

Piezoelectricity of the Transmembrane Protein ba3 Cytochrome c Oxidase

Controlling the electromechanical response of piezoelectric biological structures including tissues, peptides, and amino acids provides new applications for biocompatible, sustainable materials in electronics and medicine. Here, the piezoelectric effect is revealed in another class of biological materials, with robust longitudinal and shear piezoelectricity measured in single crystals of the transmembrane protein ba₃ cytochrome c oxidase from Thermus thermophilus. The experimental findings from piezoresponse force microscopy are substantiated using a range of control measurements and molecular models. The observed longitudinal and shear piezoelectric responses of ≈ 2 and 8 pm V⁻¹, respectively, are comparable to or exceed the performance of commonly used inorganic piezoelectric materials including quartz, aluminum nitride, and zinc oxide. This suggests that transmembrane proteins may provide, in addition to physiological energy transduction, technologically useful piezoelectric material derived entirely from nature. Membrane proteins could extend the range of rationally designed biopiezoelectric materials far beyond the minimalistic peptide motifs currently used in miniaturized energy harvesters, and the finding of robust piezoelectric response in a transmembrane protein also raises fundamental questions regarding the molecular evolution, activation, and role of regulatory proteins in the cellular nanomachinery, indicating that piezoelectricity might be important for fundamental physiological processes.     

Lithiation MXene Derivative Skeletons for Wide-Temperature Lithium Metal Anodes

Lithium (Li) metal, as an appealing candidate for the next-generation of high-energy-density batteries, is plagued by its safety issue mainly caused by uncontrolled dendrite growth and infinite volume expansion. Developing new materials that can improve the performance of Li-metal anode is one of the urgent tasks. Herein, a new MXene derivative containing pure rutile TiO₂ and N-doped carbon prepared by heat-treating MXene under a mixing gas, exhibiting high chemical activity in molten Li, is reported. The lithiation MXene derivative with a hybrid of LiTiO₂-Li₃N-C and Li offers outstanding electrochemical properties. The symmetrical cell assembling lithiation MXene derivative hybrid anode exhibits an ultra-long cycle lifespan of 2000 h with an overpotential of ≈30 mV at 1 mA cm⁻², which overwhelms Li-based anodes reported so far. Additionally, long-term operations of 34, 350, and 500 h at 10 mA cm⁻² can be achieved in symmetrical cells at temperatures of −10, 25, and 50 °C, respectively. Both experimental tests and density functional theory calculations confirm that the LiTiO₂-Li₃N-C skeleton serves as a promising host for Li infusion by alleviating volume variation. Simultaneously, the superlithiophilic interphase of Li₃N guides Li deposition along the LiTiO₂-Li₃N-C skeleton to avoid dendrite growth.     

基于IP 模块的片上系统设计

传统IC设计方法关注的是如何创建一个全新的设计并进行有效的验证。复杂的IP模块，嵌入式软件，不断增长的晶体管数量，这些都变成了传统方法日益沉重的负担，在片上系统SOC（system-on-chip)设计中，基于IP模块的功能组装正在逐渐代替传统的功能设计而成为主流的设计方法，本文介绍基于IP模块的片上系统的设计和验证。     

应用ANN的数字信号调制模式识别的特征矢量分析

数字信号调制模式的自动识别是无线通信中信号分析和处理的关键步骤之一。神经网络应用于数字信号调制模式识别的一个重要的预处理过程就是要形成完备的特征矢量空间。本文在原有分析方法的基础上，研究了信号的瞬时特征、谱特征和统计特征，特别提出了用于信号子类分析的新的高阶统计量特征，形成了数字调制模式自动识别的较为完备的特征矢量空间。通过对九种数字调制信号进行仿真表明，即使在信噪为5dB的条件下仍然能够获得较高的识别正确率。     

电容式高分子湿敏元件的研究及进展

介绍了电容式高分子湿敏元件的几个组成部分一下电极、湿敏功能材料、上电极工艺制造技术。同时也介绍了湿敏元件防污染、电路一体化设计、长期稳定性等方面近来的研究成果。指出湿敏元件的感湿材料的选择和上电极工艺制造技术是影响湿敏元件性能的两个重要因素。     

光纤传输网综合网络管理系统

本给出光纤传输网网管系统的目标、特点和体系结构,并具体给出了光纤传输网网管系统功能需求。     

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。