新型混合脂质体的制备及其与药物的分子识别研究

设计、合成了含核酸碱基的双亲化合物N6-十四酰基-9-辛基腺嘌呤三甲基溴化铵;用目标化合物和卵磷脂制备混合脂质体受体。分别研究了受体与巴比妥酸、核黄素药物在混合脂质体亲脂区通过互补氢键进行的分子识别。紫外可见光谱研究表明,底物巴比妥酸在258nm的吸收峰,核黄素在222nm,264nm的吸收峰随时间的延长而逐渐降低,这种减色作用归因于脂质体内的巴比妥酸和核黄素在亲脂区分别与双亲化合物上的腺嘌呤形成了氢键。     

基于模糊聚类的语音识别

指出模糊聚类方法影响语间识别的正确率 ,提出用基于分解构造的模糊聚类方法设立语音模式的参考向量 ,说明该方法能提高语章识别的正确率。     

Electronic tongue for quality assessment of ethanol, vodka and eau-de-vie

Rapid quality assessment of alcoholic beverages, including brand identification and detection of products of unacceptable quality or counterfeits is an important practical task. In the present work the multisensor electronic tongue system (ET), based on array of potentiometric chemical sensors was applied to recognition and classification of spirits such as vodka and ethanol used for vodka production and also for eau-de-vie in cognac production. The ET system was capable of detecting presence of contaminant substances in vodka in concentrations exceeding allowed levels as well as of distinguishing vodka complying and not complying with state quality standards. Ten brands of vodka produced at the same distillery using water and ethanol of different purity and various taste additives were discriminated using the instrument. The ET could distinguish synthetic and alimentary grain ethanol as well as alimentary ethanol of different grades (i.e. different degree of purification). A feasibility study was run on several eau-de-vie samples, which included fresh and aged eau-de-vie as well as samples produced using different distillation technology and samples kept in contact with different kinds of oak. The electronic tongue showed a promise as an analytical instrument for rapid quality assessment of spirits.     

2-(苯并噻唑-2-基)-5-(N,N-二乙基氨基)苯酚衍生物的合成及其对H2S的识别

以4-N,N-二乙基氨基水杨醛为原料,制备了2-(苯并噻唑-2-基)-5-(N,N-二乙基氨基)苯酚衍生物(探针L),并对其结构进行了表征。在DMSO/PBS(体积比3∶7,pH=7.4)溶液中,探针L具有高选择性并可荧光"关-开"识别H_2S,在365nm紫外灯照射下,由无荧光变成蓝色荧光。实验表明,探针L识别H_2S的检测限为2.05×10~(-6)mol/L,pH适用范围为6～9,可用于检测实际水样中的H_2S。     

镀金层识别及其厚度测定的X—荧光强度比值法

利用Ａｕ原子在Ｘ射线激发下所发射的Ｌβ线和Ｍ线的强度比值识别样品是镀金还是Ｋ金，同时利用这个比值测定镀金层的厚度，镀金层厚度的测定范围为０～４μｍ，测定值与标定值的相对误差小于１５％。     

基于轻量化深度学习网络的数字仪表识别

数字式仪表常用于变电站、工厂等生产环境,是一种直观的设备监测仪器。然而当前数字式仪表的读取方式还依赖于人工巡检,手动记录等,这些传统的巡检方式来监测设备的运行状态大大降低了巡检效率。为了实现传统行业的数字化转型,本文提出基于轻量化深度学习的数字仪表识别方法,通过改进的YOLOv5的目标检测框架,针对数字仪表目标区域在整张图片大小不一致的情况,提出对于感兴趣区域（ROI）的迭代目标检测方法,首次检测将感兴趣区域进行检测并切割统一到相同的尺度,随后迭代检测网络针对感兴趣区域内的字符进行检测并分类,以达到精确读数的目的。为提升多尺度检测性能,本文采用Res2Net模块主干网络中的的残差模块。采用GIoU取代通用的IoU作为位置损失函数加速模型训练效果的收敛。实验表明,改进后的框架实现了99.62%的准确率和99.55%的召回率,相比基线网络分别提升了12.72%和5.85%。通过将框架在边缘计算平台上的终端部署,在实际生产中取代了人工巡检,实现了商业化运行。     

A selective fluorescent probe for La and Y based on calix[6]arene

A new fluorescent probe based on calix[6]arene functionalized with three naphthoic acid groups was synthesized and showed selective fluorescence enhancement in the presence of La³⁺ and Y³⁺. In addition, the fluorescence enhancement behaviors depended on the pH values of the solution.     

Development of calixarene-based host molecules for peptides in aqueous media

We synthesized a peptidocalix[4]arene library consisting of 1000 members that was suitable for peptide recognition in aqueous media. Some peptidocalix[4]arenes in the library were host molecules for guest peptides. Electrostatic interaction between the host and guest molecules was the most important factor for binding in aqueous media.     

基于特征脸的人脸检测与识别

人脸识别是属于生物特征识别范畴.然而与其他识别不同的是人脸是一个动态的图像,有位置与光照的变化,并且具有相似的外貌特征,这使得人脸识别成为一个比较困难的问题.文章阐述了基于特征脸的人脸检测与识别方法.该方法本质上是主要成分分析方法,他首先构造特征脸向量,然后计算新图像和数据库中特征脸的空间距离,来决定此图像是否是一副人脸图像,如果是人脸图像,他是那一副人脸图像。     

《生本数学》:一种全新的课程观

《生本数学》将知识分"感受""认识""熟悉"三个层次进行学习,主张知识的整合,其课程的设置结构合理,循序渐进,符合学生的学习规律,因而是一种全新的课程观。