在线超临界流体萃取-超高效液相色谱-光电二极管检测器联用分析胡椒粉中的胡椒碱

本文介绍的联用分析系统是将超临界流体萃取（SFE）系统与超高效液相色谱（UHPLC）系统联用。使用在线SFE-UHPLC系统,可以完成对胡椒粉中胡椒碱成分的高速萃取和分析。整个分析过程包括萃取时间在内只需要7min。将此方法的萃取回收率与其他萃取方法进行比较,SFE方法得到的回收率与其他方法相当。这种在线SFE-UHPLC方法操作简单,分析速度快。这一方法对于分析固体样品中的化合物组分非常有用,可以有效避免繁琐的样品预处理过程。     

Effects of exogenous mevalonic acid on sterol lipid classes in Larix kaempferi callus

In Japanese larch (Larix kaempferi (Lamb.) Carr.) calli, free sterol (FS), acylsterol (AS) and glycosylsterol, including the acylated type, were found in the proportion of 1.0:0.1:0.8. When the calli were cultured in the presence of 10 mM mevalonic acid (MVA), the content of AS, but not FS and glycosylsterol, was increased remarkably. The major component sterol in each sterol lipid class was usually sitosterol (more than 90%) with campesterol as a minor one. There were no differences on the sterol compositions between the calli cultured with or without MVA. When the calli cultured with 10 mM MVA for 6 weeks were transferred to the control medium without exogenous MVA, AS contents decreased to the level of the control calli. Thus, it was shown that sterol lipids, such as FS and glycosylsterols, with the structural functions was maintained in the constant content and the excess sterol biosynthesized from exogenous MVA was esterified to form AS for storage of sterol components.     

Fabrication of multilayered photochromic memory media using pressure-sensitive adhesives

We fabricated a multilayered medium by a laminating process with pressure-sensitive adhesives. It was possible to reduce the thickness variations of both photosensitive layers and transparent layers by applying laminating films. This method is easy to use to pile up many layers for a multilayered medium. We fabricated twenty recording layers and demonstrated the capability to record with significant reading results. The recorded data in each layer were read out without crosstalk. We evaluated the signal-to-noise ratio and crosstalk between neighboring layers. It was found that the signal-to-noise ratio of a multilayered medium was higher than 50 dB.     

Rapid enumeration of low numbers of moulds in tea based drinks using an automated system

Aseptically prepared cold drinks based on tea have become popular worldwide. Contamination of these drinks with harmful microbes is a potential health problem because such drinks are kept free from preservatives to maximize aroma and flavour. Heat-tolerant conidia and ascospores of fungi can survive pasteurization, and need to be detected as quickly as possible. We were able to rapidly and accurately detect low numbers of conidia and ascospores in tea-based drinks using fluorescent staining followed by an automated counting system. Conidia or ascospores were inoculated into green tea and oolong tea, and samples were immediately filtered through nitrocellulose membranes (pore size: 0.8 μm) to concentrate fungal propagules. These were transferred onto potato dextrose agar and incubated for 23 h at 28 °C. Fungi germinating on the membranes were fluorescently stained for 30 min. The stained mycelia were counted selectively within 90 s using an automated counting system (MGS-10LD; Chuo Electric Works, Osaka, Japan). Very low numbers (1 CFU/100 ml) of conidia or ascospores could be rapidly counted, in contrast to traditional labour intensive techniques. All tested mould strains were detected within 24 h while conventional plate counting required 72 h for colony enumeration. Counts of slow-growing fungi (Cladosporium cladosporioides) obtained by automated counting and by conventional plate counting were close (r² = 0.986). Our combination of methods enables counting of both fast- and slow-growing fungi, and should be useful for microbiological quality control of tea-based and also other drinks.     

文本分类的特征提取方法比较与改进

文本的特征提取是文本分类过程中的一个重要环节,它的好坏将直接影响文本分类的准确率。该文介绍了词条的χ2统计方法（CHI）、词条与类别的互信息（MI）、信息增益（IG）、词条的期望交叉熵（CE）等文本特征提取方法,并对其取词策略进行了改进。为了对这些特征提取方法进行系统地比较,选择了三种代表性的分类器对《读卖新闻》文本数据库进行了分类实验。实验结果表明χ2统计方法具有最好的准确率,各种改进的特征提取方法都能提高文本分类的准确率。     

Attention-based neural network with Generalized Mean Pooling for cross-view geo-localization between UAV and satellite

Artificial Life and Robotics - Cross-view geo-localization is finding images containing the same geographic target in multi-views. For example, given a query image from UAV view, a proposed...