基于小波和轮廓提取的色谱基线算法研究

把基于小波和轮廓提取的基线算法引入了色谱基线提取领域;基于轮廓提取的算法为:通过构造一个滑动窗,沿着色谱数据滑动求窗内的最小值。把这些最小值插值平滑就获得了色谱基线。基于小波和轮廓提取的算法为:先用小波初步提取基线,然后把色谱数据减去小波提取的基线后再用轮廓提取算法获得基线,把小波提取的基线和轮廓提取的基线相加即为原数据的基线。对这两个算法进行了比较实验研究,结果表明:基于小波和轮廓提取的算法比轮廓提取的算法效果好,能更准确地提取色谱基线。     

疲劳驾驶检测中基于稀疏表示的眼睛状态识别研究

为了解决传统接触式疲劳驾驶检测方法影响驾驶、检测算法识别率较低等问题,本文提出一种基于稀疏表示的眼睛状态识别的方法。利用K-SVD(K均值奇异值分解)方法对输入的训练集构造过完备冗余字典,利用正交匹配追踪法对测试的图像进行稀疏表示,然后根据重构图像和测试图像之间的误差,确定测试图像所属的类别,判断出测试图像的状态。实验中将K-SVD和OMP(正交匹配追踪)方法与其它字典学习和稀疏表示方法进行对比,结果表明,利用K-SVD字典学习算法结合OMP算法获得了较好的识别效果。     

三维人脸识别算法研究

近年来,基于三维图像的人脸识别技术已经取得了很大进展,在约束环境下也能获得很好的识别性能,但仍受限于姿态、表情等因素,需要从算法上改进才能解决其影响。本文分别从基于空域直接匹配、基于局部特征匹配和基于整体特征匹配3个角度出发,对人脸匹配算法以及融合算法进行了研究,列出了部分改进算法的实验结果,并分析了算法有效性的原因,总结了目前面临的三维人脸识别算法难以突破的一些困难及未来的研究趋势。     

液相色谱-质谱用于卵巢肿瘤中磷脂轮廓的分析

卵巢肿瘤日益影响女性的健康和生活质量,其中的卵巢癌是女性三大恶性肿瘤之一,死亡率高居三者之首。因此卵巢肿瘤尤其卵巢癌是目前的一个研究热点。本研究利用液相色谱-质谱(LC-MS)联用技术对卵巢肿瘤进行磷脂轮廓分析,研究良性卵巢肿瘤(B)和卵巢癌(M)的患者血清中磷脂代谢的差异情况。首先用LC-MS采集血清中磷脂的指纹图谱,通过峰识别、峰匹配等得到峰表,然后利用正交校正的偏最小二乘法(OSC-PLS)进行多种分型,根据模型的变量重要因子(VIP)、VIP值的置信区间、S图和显著性差异检验结果等筛选有差异的磷脂。结果显示: M组和B组与正常对照(N)组比较都存在明显的磷脂代谢差异,发生改变的磷脂主要为缩醛磷脂酰乙醇胺、磷脂酰胆碱、缩醛磷脂酰胆碱、鞘磷脂和溶血磷脂酰胆碱。     

基于液相色谱与质谱联用的代谢组学及磷脂轮廓分析在糖代谢异常研究中的应用

糖代谢异常由于其发病率的升高和影响人类的生活质量而日益受到科学研究者的关注。实验中利用液相色谱与质谱(LC-MS)联用技术对糖代谢异常分别进行了代谢组学和磷脂轮廓分析,研究了糖代谢异常中的两个阶段——空腹血糖受损(IFG)和初诊糖尿病(NDD)的代谢差异情况。首先从LC-MS采集到血浆中代谢组学分析及磷脂轮廓分析的原始谱图,通过软件的峰匹配等步骤得到峰表,之后利用多种统计分析方法进行数据分析,通过正交校正的偏最小二乘法(OSC-PLS)对样品进行分型,根据模型的变量重要因子(VIP)、显著性差异检验结果等筛选出差异性代谢物。结果显示: NDD组比IFG组与对照组(N组)比较存在更明显的代谢差异,发生变化的化合物主要为游离脂肪酸、溶血磷脂酰胆碱、磷脂酰乙醇胺、鞘磷脂和磷脂酰胆碱等。     

基于高光谱技术的培养基上细菌菌落分类方法研究

利用高光谱技术对培养基上细菌(大肠杆菌、李斯特菌和金黄色葡萄球菌)菌落进行快速识别和分类。采集琼脂培养基上细菌菌落的高光谱反射图像(390~1040 nm),在对波段差图像进行大津阈值分割的基础上自动提取细菌菌落光谱,并建立细菌分类检测的全波长和简化偏最小二乘判别( PLS-DA)模型。全波长模型对预测集样本的分类准确率和置信预测分类准确率分别为100%和95.9%。此外,利用竞争性自适应重加权算法( CARS)、遗传算法( GA)和最小角回归算法( LARS-Lasso)进行波长优选并建立对应简化模型。其中,CARS简化模型在精度、稳定性及分类准确率方面均优于GA和LARS-Lasso简化模型,其对预测集样本的分类准确率和置信预测分类准确率分别达到了100%和98.0%。研究表明,高光谱是一种细菌菌落高精度、快速、无损识别检测的有效方法。简化模型中优选的波长可以为开发低成本检测仪器提供理论依据。     

基于特征块匹配的医用注射液图像位移补偿应用

针对机器视觉药液异物检测机器人由于机械振动或跟踪不同步,导致所拍摄的序列图像产生位移偏差、影响后续异物分割与提取的问题,本工作采取特征点匹配与块匹配相结合的方法,对序列图像进行配准,求出运动矢量进行图像位移补偿,克服了传统特征点匹配产生空间位置偏差较大误匹配、传统块匹配需对背景静止块进行预处理以及单独使用特征点对匹配或块匹配均难以满足实时需要的问题。首先,对参考帧进行FAST特征点检测,再利用Hessian矩阵、非极大值抑制、熵值差法和特征点距离约束选取稳定的特征点;然后,根据特征点位置信息,选取以特征点为中心的待匹配宏块,再采用一种改进的偏水平方向的六边形搜索策略（HHS）与当前帧进行块与块的匹配,找到最佳匹配块;最后,利用匹配块之间的坐标参数求出运动参数,再利用求出的运动矢量进行帧间补偿。实验结果表明,该算法实时性能能达到72 ms,远快于点对匹配中ORB算法的140 ms,比直接用原有的六边形搜索算法（HS）进行块匹配快了近20%,兼顾了速度与精度,能快速补偿药瓶在图像中的位移偏差,抑制图像位移偏差所引起的各种干扰。     

液相色谱-质谱联用系统在肝硬化不同阶段代谢轮廓研究中的应用

采用高效液相色谱-轨道离子阱质谱联用(HPLC-LTQ Orbitrap XL MS)代谢组学研究平台分析不同阶段肝硬化病人和健康人群的血清标本,获取代谢轮廓。采用模式识别方法结合非参数检验对数据进行分析。研究发现,由肝硬化A级组、B级组、C级组和健康对照组的代谢轮廓构建的正交偏最小二乘判别分析(OPLS-DA)模型(R2（Y）=90.1%, Q2=66.7%),对检测组数据的预测准确率达到93.8%,具有很好的判别能力。从代谢轮廓中可以鉴别出用于区分不同疾病阶段的特异性代谢标志物,如溶血磷脂酰胆碱、甘氨鹅去氧胆酸、半胱氨酸、甘氨酸、氨基己二酸、哌可酸等。研究结果表明: 利用代谢组学方法获得的血清代谢轮廓可以用来构建区分模型和寻找代谢标志物,为乙肝肝硬化的诊断和监测提供支持和依据。     

基于霍夫变换的球状TEM图像的识别方法

介绍霍夫变换原理以及球状图像的检测过程，利用Matlab软件设计程序识别球状TEM图像，通过实例说明，利用霍夫变换找圆算法可以实现对TEM图像中圆形粒子大小以及位置进行识别。     

一种基于块模糊增强的玉米田遥感图像边缘检测方法

针对高分一号卫星(GF-1)玉米田遥感图像中玉米田光谱复杂和地块边缘模糊导致的面积统计误差大的问题,本文提出一种块模糊增强和最小值边缘提取相结合的边缘检测方法进行玉米田地块分割处理,以减小面积统计误差。首先将彩色遥感图像从RGB变换到I¹I²I³彩色空间,提取出含丰富特征的单色图I¹;然后利用模糊理论对I¹进行基于块的增强处理;再对增强后的图像进行最小值边缘提取;最后利用Full Lambda-Schedule算法对区域边缘进行优化。通过与Canny和Sobel等边缘提取方法比较,证明本文的边缘检测结果能有效地分割出玉米田地块目标,减少了玉米田光谱复杂和边缘模糊带来的影响,检测出的边缘更符合玉米田实际分布,玉米田面积统计结果更符合实际。     

A smartphone‐based on‐site nucleic acid testing platform at point‐of‐care settings

We developed a smartphone‐based on‐site nucleic acid testing (NAT) platform that can image and analyze lateral flow nucleic acid assays at point‐of‐care settings. An inexpensive add‐on was devised to run lateral flow assays while providing homogeneous ambient light for imaging. In addition, an Android app with a user‐friendly interface was developed for the result analysis and management. Linear color calibration is implemented inside the app to minimize the colorimetric reaction difference between smartphones. A relationship function between nucleic acid concentration and colorimetric reaction was established and evaluated by leave‐one‐out cross validation. The predicted concentration and true concentration showed a good agreement with an R‐squared value of 0.96. This smartphone‐based NAT platform can be used to diagnose infectious diseases and monitor disease progression, and assess treatment efficacy, especially for resource‐limited settings.     

Detailed Visualization of Aromaticity Using Isotropic Magnetic Shielding

For many years, Clar's aromatic sextet theory has served as a qualitative method for assessing the aromatic character of polycyclic aromatic hydrocarbons. A new approach, based on the calculation of isotropic magnetic shielding (IMS) contour plots, is shown to provide a feature-rich picture of aromaticity that is both quantitative yet still easily interpreted. Chemists are visual creatures who are adept at discerning reactivity and chemical behavior from molecular structures. To quote Roald Hoffmann, “People like pictures. Chemists live off them.” Thus, the detailed image analysis we present simultaneously provides quantitative assessment of electronic structure, which is still easy-to-understand through visual inspection, embedded in an aesthetically appealing and intuitive picture that draws the reader in. We provide novel computed IMS contour plots for a representative selection of aromatic molecules. Where Clar's static drawings capture only a partial sketch of the electronic properties of a molecule, IMS contour plots present a detailed, global landscape of a molecule that sums all possible resonance structures. This novel analysis allows us to correct certain drawbacks of Clar's analysis with respect to polycyclic aromatics and quantitatively assess the bonding and electronic structure of acene hydrocarbons.     

Smartphone‐Based Android app for Determining UVA Aerosol Optical Depth and Direct Solar Irradiances

This research describes the development and evaluation of the accuracy and precision of an Android app specifically designed, written and installed on a smartphone for detecting and quantifying incident solar UVA radiation and subsequently, aerosol optical depth at 340 and 380 nm. Earlier studies demonstrated that a smartphone image sensor can detect UVA radiation and the responsivity can be calibrated to measured direct solar irradiance. This current research provides the data collection, calibration, processing, calculations and display all on a smartphone. A very strong coefficient of determination of 0.98 was achieved when the digital response was recalibrated and compared to the Microtops sun photometer direct UVA irradiance observations. The mean percentage discrepancy for derived direct solar irradiance was only 4% and 6% for observations at 380 and 340 nm, respectively, lessening with decreasing solar zenith angle. An 8% mean percent difference discrepancy was observed when comparing aerosol optical depth, also decreasing as solar zenith angle decreases. The results indicate that a specifically designed Android app linking and using a smartphone image sensor, calendar and clock, with additional external narrow bandpass and neutral density filters can be used as a field sensor to evaluate both direct solar UVA irradiance and low aerosol optical depths for areas with low aerosol loads.     

中药注射剂三维荧光等高线图识别程序的设计与检验

用统计模式识别法设计了中药注射剂三维荧光指纹图谱的识别程序,对刺五加、红花、血栓通和香丹4种注射剂进行了识别,识别结果正确。研究表明:该程序分辨率高,便于自动化,可用于中药注射剂的真伪鉴别和质量监控。     

亚硫酸盐对彩色成像过程的影响

研究了彩色显影液中亚硫酸钠对含内偶青成色剂的卤化银单层片的彩色成像过程的各化学反应、青染料影像的有害吸收以及染料云形态的影响,结果表明亚硫酸钠能(1)提高银影像的形成速度;(2)降低成色剂的表观偶合效率;(3)增加青染料影像的有害吸收;(4)使染料云缩小并且使染料云的分布变窄。     

Developing a mobile app for remote access to and data analysis of spectra

Vibrational spectroscopy is a non-destructive analytical method that can be used to analyse a wide range of materials. A vibrational spectrum contains information on the chemical structure of the sample being analysed, which can be recorded rapidly. With hand held mobile device technology being considered as a relatively mature market, there is an excellent opportunity to combine vibrational spectroscopy with mobile devices for in situ analysis of samples. There are still instances where analytical instruments require being linked to desktop PC's/laptops for instrument control and data manipulation. However, mobile devices are becoming increasingly more powerful thus, enabling data manipulation on devices via cloud based technology. With desktop PC's and laptops often having a larger environment footprint than the instrumental spectrometer itself, this therefore highlights the potential for mobile spectroscopy devices. This paper reports the first development of an app (SpectralAnalyser) to enable the use of mobile devices to access and manipulate spectra and describes the different approaches and implementation issues considered during the development of apps to display spectra on Android and iOS platforms.     

T型微混合器混合特性的浓度分布评价法

提出一种微混合器混合性能的评价方法.在样品盒中注入不同浓度罗丹明B溶液并用体视显微镜观察捕获图像,通过Image J软件读取图像灰度值,建立不同深度下溶液的浓度-灰度值函数关系,运用此关系式将T型微混合器3种不同深度(0.1、 0.2和0.4 mm,质量浓度0.05 %的罗丹明B溶液和去离子水作为配对流体)混合实验中捕获的图像中各像素点上的灰度值转换为浓度值,绘制浓度等高线图及浓度频数分布图,分析各自混合情况,最后引入浓度混合指数概念及计算公式,分析3种深度混合器内不同截面上的混合程度.此方法从定性和定量两方面分析了微尺度下混合腔深度对微混合的影响程度,具有一定的应用价值.     

Optimization of injection‐molding process for mechanical properties of polypropylene components via a generalized regression neural network

This study analyzed contour distortions, wear and tensile properties of polypropylene (PP) components applied in the interior coffer of automobiles. A hybrid method integrating a trained generalized regression neural network (GRNN) and a sequential quadratic programming (SQP) method is proposed to determine an optimal parameter setting of the injection‐molding process. The specimens were prepared under different injection‐molding conditions by changing melting temperatures, injection speeds, and injection pressures. Average contour distortions at six critical locations, wear and tensile properties were selected as the quality targets. Sixteen experimental runs, based on a Taguchi orthogonal array table, were utilized to train the GRNN and then the SQP method was applied to search for an optimal setting. The trained GRNN was capable of predicting average contour distortions, wear and tensile properties at various injection‐molding conditions. In addition, the analysis of variance (ANOVA) was implemented to identify significant factors for the molding process and the proposed algorithm was compared with traditional schemes like the Taguchi method and the design of experiments (DOE) approach. Copyright © 2007 John Wiley & Sons, Ltd.     

Optimal cloud use of quartic force fields: The first purely commercial cloud computing based study for rovibrational analysis of SiCH−

Commercial cloud computing (CCC) has the promise of an untold number of computing nodes available for the researcher as long as he or she has the financial means to absorb these costs and the administrative skills necessary to effectively utilize the resources. The key is finding how to maximize parallelization for a minimum of monetary and management costs. Previous work has shown that CCC resources are viable for use on large numbers of small‐to‐medium sized quantum chemical computations. Composite energy quartic force fields (QFFs) are a highly‐attractive platform for subsequent testing of CCC resources to find the proper balance between time savings of the cloud versus monetary expenditure. Use of this type of potential energy surface has lead to highly‐accurate rovibrational data in earlier work. QFFs use large numbers of stand‐alone energies that have to be computed for various molecular geometries. At each geometry, different methods and/or basis sets are used to efficiently generate accurate representations of the nuclear potential. For this initial study, the small molecular anion, SiCH^? of interest in astrochemistry, is chosen for analysis as it can be done cheaply on the cloud while still providing insight into the nature of CCC usage. Additionally, no rovibrational data exists for this molecule making it the first molecule quantum chemically computed purely via CCC tools. © 2015 Wiley Periodicals, Inc.     

Synthesis of peptide alkylthioesters using the intramolecular N,S-acyl shift properties of bis(2-sulfanylethyl)amido peptides

The design of novel methods giving access to peptide alkylthioesters, the key building blocks for protein synthesis using Native Chemical Ligation, is an important area of research. Bis(2-sulfanylethyl)amido peptides (SEA peptides) 1 equilibrate in aqueous solution with S-2-(2-mercaptoethylamino)ethyl thioester peptides 2 through an N,S-acyl shift mechanism. HPLC was used to study the rate of equilibration for different C-terminal amino acids and the position of equilibrium as a function of pH. We show also that thioester form 2 can participate efficiently in a thiol-thioester exchange reaction with 5% aqueous 3-mercaptopropionic acid. The highest reaction rate was obtained at pH 4. These experimental conditions are significantly less acidic than those reported in the past for related systems. The method was validated with the synthesis of a 24-mer peptide thioester. Consequently, SEA peptides 1 constitute a powerful platform for access to native chemical ligation methodologies.