Study on the Cross-Cultural Identification of A Community of Shared Future for Mankind

As a kind of“global view”beyond the country and ideology,the concept of a community of shared future for mankind is the core of Xi Jinping’s diplomatic thoughts,and also a Chinese strategy for the future of mankind,which is proposed by China to adapt to the interrelation of human societies with a view of world peace and development.The cross-cultural awareness contained in such concept refers actually to the promotion and abandonment of different cultures in the world.Cross-cultural identification means that the national culture constantly absorbs the positive factors of foreign cultures while discards its negative factors,so as to promote the benign combination of the both and realize the localization of the latter.In the longer term,the most direct mission of the proposal of such concept is to answer the goal of the great rejuvenation of the Chinese nation,and to promote the common revitalization of all countries and resurgence of civilizations.This concept interprets the relationship between China and the world and also integrates the Chinese dream with the world dream.Therefore,the study on cross-cultural identification of all countries should expand international understanding and broad consensus,on the basis of the global perspective of the community of shared future for mankind,and under the principle of“symbiosis,negotiation,collaborative construction,sharing and a win-win situation”.

Study on Identification and Fermentation Condition of Strain ZY-19-2 with Inhibitory Effect against Tobacco Black Shank (Phytophtora parasitica var. nicotianae Tucker)

[Objective] The paper was to identify stain ZY-19-2 with inhibitory effect against tobacco black shank （Phytophtora parasitica var.nicotianae Tucker）, and study the fermentation condition of the strain. [Method]A strain ZY-19-2 with strong inhibitory effect against P. parasitica were isolated and screened from tobacco rhizosphere soil samples, and identified according to its morphological characteristics. The chitinase production activity of the strain under different culture conditions was also studied. [Result] For stain ZY-19-2 Paecilomyces lilacinus, the optimal fermentation conditions were as follows： 1.2% colloidal chitin as carbon source, 1% peptone as nitrogen source, 0.1% Tween 80 as surfactant, initial pH of fermentation broth at 6.0, the fermentation time of 60 h, inoculum amount at 1%, shaker speed at 120 r/min. The highest enzyme activity reached 0.216 U/ml. [Conclusion]The optimization of fermentation condition of strain ZY-19-2 lay foundation for large-scale production of cheap and efficient chitinase and chitin oligosaccharides, as well as application of the strain for control of tobacco black shank.

Modal identification based on Hilbert-Huang Transform of structural response with S VD preprocessing

In recent years, Empirical mode decomposition and Hilbert spectral analysis have been combined to identify system parameters. Singular-Value Decomposition is pro- posed as a signal preprocessing technique of Hilbert-Huang Transform to extract modal parameters for closely spaced modes and low-energy components. The proposed method is applied to a simulated airplane model built in Automatic Dynamic Analysis of Mechanical Systems software. The results demonstrate that the identified modal parameters are in good agreement with the baseline model.

An Insect Imaging System to Automate Rice Light-Trap Pest Identification

Identification and counting of rice light-trap pests are important to monitor rice pest population dynamics and make pest forecast. Identification and counting of rice light-trap pests manually is time-consuming, and leads to fatigue and an increase in the error rate. A rice light-trap insect imaging system is developed to automate rice pest identification. This system can capture the top and bottom images of each insect by two cameras to obtain more image features. A method is proposed for removing the background by color difference of two images with pests and non-pests. 156 features including color, shape and texture features of each pest are extracted into an support vector machine （SVM） classifier with radial basis kernel function. The seven-fold cross-validation is used to improve the accurate rate of pest identification. Four species of Lepidoptera rice pests are tested and achieved 97.5% average accurate rate.

Transformer-CNN特征跨注意力融合学习的行人重识别

卷积神经网络(convolutional neural network,CNN)关注局部特征,难以获得全局结构信息,Transformer网络建模长距离的特征依赖,但易忽略局部特征细节。提出了一种跨注意力融合学习的行人重识别算法,利用CNN和Transformer特征学习网络的特点,在丰富行人局部特征的同时改善特征的全局表达能力。该模型由三个部分构成:CNN分支主要提取局部细节信息;Transformer分支侧重于关注全局特征信息;跨注意力融合分支通过自注意力机制计算上述两个分支特征的相关性,进而实现特征融合,最终提高模型的表征能力。剥离实验以及在Market1501和DukeMTMC-reID数据集的实验结果证明了所提方法的有效性。

四流输入引导的特征互补可见光-红外行人重识别

目前可见光-红外行人重识别研究侧重于通过注意力机制提取模态共享显著性特征来最小化模态差异。然而,这类方法仅关注行人最显著特征,无法充分利用模态信息。针对此问题,本文提出了一种四流输入引导的特征互补网络(QFCNet)。首先在模态特定特征提取阶段设计了四流特征提取和融合模块,通过增加两流输入,缓解模态间颜色差异,丰富模态的语义信息,进一步促进多维特征融合;其次设计了一个次显著特征互补模块,通过反转操作补充全局特征中被注意力机制忽略的行人细节信息,强化行人鉴别性特征。在SYSU-MM01,Reg DB两个公开数据集上的实验数据表明了此方法的先进性,其中在SYSU-MM01的全搜索模式中rank-1和m AP值达到了76.12%和71.51%。

基于跨模态近邻损失的可视-红外行人重识别

可视-红外跨模态行人重识别任务的目标是给定一个模态的特定人员图像,在其他不同模态摄像机所拍摄的图像集中进行检索,找出相同人员对应的图像。由于成像方式不同,不同模态的图像之间存在明显的模态差异。为此,从度量学习的角度出发,对损失函数进行改进以获取具有更加辨别性的信息。对图像特征内聚性进行理论分析,并在此基础上提出一种基于内聚性分析和跨模态近邻损失函数的重识别方法,以加强不同模态样本的内聚性。将跨模态困难样本的相似性度量问题转化为跨模态最近邻样本对和同模态样本对的相似性度量,使得网络对模态内聚性的优化更加高效和稳定。对所提方法在全局特征表示的基线网络和部分特征表示的基线网络上进行实验验证结果表明:所提方法对可视-红外行人重识别的预测结果相较于基线方法,平均准确度最高可提升8.44%,证明了方法在不同网络架构中的通用性;同时,以较小的模型复杂度和较低的计算量为代价,实现了可靠的跨模态行人重识别结果。

Inverse Load Identification in Stiffened Plate Structure Based on in situ Strain Measurement

For practical engineering structures,it is usually difficult to measure external load distribution in a direct manner,which makes inverse load identification important.Specifically,load identification is a typical inverse problem,for which the models(e.g.,response matrix)are often ill-posed,resulting in degraded accuracy and impaired noise immunity of load identification.This study aims at identifying external loads in a stiffened plate structure,through comparing the effectiveness of different methods for parameter selection in regulation problems,including the Generalized Cross Validation(GCV)method,the Ordinary Cross Validation method and the truncated singular value decomposition method.With demonstrated high accuracy,the GCV method is used to identify concentrated loads in three different directions(e.g.,vertical,lateral and longitudinal)exerted on a stiffened plate.The results show that the GCV method is able to effectively identify multi-source static loads,with relative errors less than 5%.Moreover,under the situation of swept frequency excitation,when the excitation frequency is near the natural frequency of the structure,the GCV method can achieve much higher accuracy compared with direct inversion.At other excitation frequencies,the average recognition error of the GCV method load identification less than 10%.

Estimation of Cloud Motion Using Cross-Correlation

This paper describes the estimation of cloud motion using lag cross-correlation. In order to compute the lag cross correlation, the Bayes Decision method is used first to identify cloud and surface of earth. Then cloud motion vectors are retrieved at a subset of points through multiple applications of a cross-correlation analysis. An objective analysis is used to define displacement at every satellite pixel throughout the domain and smooth the local inconsistencies. Cloud motions are then produced with a backward trajectory technique by using these displacement vectors.

RESEARCH ON THE ALGORITHM OF BINARY IMAGE CROSS-CORRELATION FOR UNSTEADY FLOW FIELD MEASUREMENT

The algorithm of Binary Image Cross-Correlation (BICC) was developed to measure the unsteady flow field. A vortex flow field was used to test the algorithm by numerical simulation. The results show that BICC is an effective algorithm for particle identification from consecutive images, the accurate velocity vector field can be obtained. The real velocity field in a valve chamber was measured by BICC in this study. From the full-field velocity information, the pressure and vorticity fields were also extracted by post-processing. (Edited author abstract) 6 Refs.

USING CROSS-CORRELATION THEORY TO EXTRACT MODAL PARAMETERS IN FREQUENCY- DOMAIN

Conventional modal parameter identifications are usually based on frequencyresponse functions, which require measurements of both the input force and the resulting response.However, in many cases, only response data are available while the actual excitations (such aswind/wave load) are not measurable. Modal parameters estimation must base itself on response-onlydata. Over the past years, many time-domain modal parameter identification techniques fromoutput-only are proposed. A poly-reference frequency-domain modal identification scheme onresponse-only is presented. It is based on coupling the cross-correlation theory with conventionalfrequency-domain modal parameter extraction. An experiment using an airplane model is performed toverify the proposed method.

Study on the Microscopic Identification of Chenopodium album L.

[Objectives]This study was conducted to identify the microscopic characteristics of Chenopodium album L.[Methods]The microscopic identification method was adopted.[Results]The xylem vessels and fiber bundles of the roots are arranged into 3-4 intermittent ring belts in a concave-convex pattern,alternately with the parenchymal cell ring belts;and the xylem rays vary in width.The cross section of the stems is polygonal;there are parenchyma cells at the corners of the cortex;there are many vascular bundles of varying sizes;and calcium oxalate clusters are common in the medulla and cortex.There are 2 to 4 vascular bundles in the main leaf vein,peltate;and fiber bundles exist below the phloem.The powder is characterized by numerous clusters of calcium oxalate,which are uniform in size,sharp at edges and corners.[Conclusions]This study can provide reference for the identification and quality standard of the crude drug.

面向可见光-红外图像的跨模态行人再识别方法

为降低模型对图像颜色的敏感度,减小可见光和红外模态间的差异,提出一种面向可见光-红外图像的跨模态行人再识别方法。将可见光图像转换到HSV颜色空间,提取只描述图像明暗信息的V分量,降低模型对颜色信息的依赖性;通过轻量级网络对V分量图像进行降维和升维,生成介于可见光和红外图像的中间模态,缩小模态间的差异性;在SYSU-MM01和RegDB数据集上进行性能评估。性能评估结果为Rank-1的数值分别增加了6.67%、1.18%,mAP的数值分别增加了6.47%、1.15%,mINP的数值分别增加了5.59%、0.42%。

一种单阶段无监督可见光-红外跨模态行人重识别方法

无监督“可见光-红外”跨模态行人重识别任务能够缓解智能监控场景中需要大量人工标注的问题。常见多阶段模型用于处理不同模态数据。文中提出了一种有效的单阶段无监督跨模态行人重识别的方法,设计了基于置信因子的聚类算法和图嵌入的跨模态特征处理方法,分别用于解决无标签问题和跨模态问题。实验结果表明,相较于现有算法,所提方法在r=1时精度至少取得了7%的提高。

Variance Reduction Technique for Estimating Value-at-Risk based on the Cross - Entropy

Value-at-Risk （VaR） estimation via Monte Carlo （MC） simulation is studied here. The variance reduction technique is proposed in order to speed up MC algorithm. The algorithm for estimating the probability of high portfolio losses （more general risk measure） based on the Cross - Entropy importance sampling is developed. This algorithm can easily be applied in any light- or heavy-tailed case without an extra adaptation. Besides, it does not loose in the performance in comparison to other known methods. A numerical study in both cases is performed and the variance reduction rate is compared with other known methods. The problem of VaR estimation using procedures for estimating the probability of high portfolio losses is also discussed.

多模态特征融合和自蒸馏的红外-可见光行人重识别

现有跨模态行人重识别方法大多挖掘模态不变的特征,忽略了不同模态内的具有判别性的自有特征.为了充分地利用不同模态内的自有特征,提出一种多模态特征融合和自蒸馏的红外-可见光行人重识别方法.首先提出一种基于双分类器的注意力融合机制,为各模态的自有特征赋予较大的融合权重,共有特征赋予较小的融合权重,得到含有各模态判别性自有特征的多模态融合特征;为了提升网络特征的鲁棒性以适应行人外观的变化,构建一个记忆存储器来存储行人的多视角特征;还设计了一种自蒸馏无参数动态引导策略,在多模态融合特征和多视角特征的引导下,利用该策略动态强化网络的多模态推理和多视角推理能力;最后网络能够从一个行人的单模态图像推理出另一模态不同视角行人特征,提升模型跨模态行人重识别的性能.基于PyTorch深度学习框架,在公开数据集SYSU-MM01和RegDB上与当前主流的方法进行对比实验,结果表明,所提方法的Rank-1分别达到63.12%和92.55%, mAP分别达到61.51%和89.55%,优于对比方法.

基于GWO-BiLSTM的岩性识别方法研究与应用

为解决常规岩性识别方法精度不高、耗时较长且受人为影响较大等问题,构建了基于GWO-BiLSTM的岩性识别方法。结合录井资料、岩心资料以及测井资料,采用常规方法对研究区进行岩性识别,效果较差,进而利用GWO-BiLSTM模型在研究区展开岩性识别工作。根据皮尔逊函数对各测井曲线与岩性进行分析,优选出相关系数绝对值大于0.3的测井曲线值作为输入特征,采用灰狼优化算法对BiLSTM超参数组合随机生成与更新,从而更加快速地获取最优解,进一步提高模型的效率以及准确率。实验表明,基于GWO-BiLSTM模型的岩性识别准确率达96%,与BiLSTM模型、RF模型、BP神经网络和SVM模型相比具有较高的准确率,验证了该模型在识别复杂岩性时的可靠性,并为复杂岩性识别提供了方法参考。

基于遗传算法-v支持向量回归的船舶轨迹预测

为了提高船舶轨迹预测精度,避免船舶海上航行事故的发生,本文采用遗传算法对v-支持向量回归进行参数寻优,以此来分别构建关于经纬度的船舶轨迹预测模型。选取水上移动业务标识码为356772000的货船在2022年6月的船舶自动识别系统数据作为研究对象。将该模型的预测结果分别与粒子群优化算法和网格搜索算法优化的v-支持向量回归模型、遗传算法-支持向量回归模型进行比较。实验结果表明:遗传算法v-支持向量回归模型关于航迹经、纬度预测结果的均方误差、平均绝对百分比误差和平均绝对误差相比于其他模型最低,关于经度分别为4.29×10^(-7)(°)、4.50×10^(-4)和5.47×10^(-7)(°)2,关于纬度的分别为1.82×10^(-6)(°)、4.02×10^(-3)和1.08×10^(-3)(°)2。基于遗传算法-v支持向量回归模型的预测效果最好,预测误差波动最小。本文将遗传算法与v-支持向量回归相结合,为船舶轨迹预测模型的优化提供参考,也为海上智能交通提供思路。

结合视觉文本匹配和图嵌入的可见光-红外行人重识别

对于可见光-红外跨模态行人重识别(Re-ID),大多数方法采用基于模态转换的策略,通过对抗网络生成图像,以此建立不同模态间的相互联系。然而这些方法往往不能有效降低模态间的差距,导致重识别性能不佳。针对此问题,该文提出一种基于视觉文本匹配和图嵌入的双阶段跨模态行人重识别方法。该方法通过上下文优化方案构建可学习文本模板,生成行人描述作为模态间的关联信息。具体而言,在第1阶段基于图片-文本对的预训练(CLIP)模型实现同一行人不同模态间的统一文本描述作为先验信息辅助降低模态差异。同时在第2阶段引入基于图嵌入的跨模态约束框架,设计模态间自适应损失函数,提升行人识别准确率。为了验证所提方法的有效性,在SYSU-MM01和Reg DB数据集上进行了大量实验,其中SYSU-MM01数据集上的首次命中(Rank-1)和平均精度均值(m AP)分别达到64.2%,60.2%。实验结果表明,该文所提方法能够提升可见光-红外跨模态行人重识别的准确率。

结缕草属植物杂交后代杂种真实性鉴定--SRAP分子标记

SRAP标记是基于PCR技术的一种新型分子标记。本研究通过SRAP分子标记技术对结缕草属植物种间杂交的5个组合37份杂交后代进行了真假杂种的鉴定,以为育种工作进一步开展奠定基础。首先筛选出在父本中能扩增出,而在母本中没有的特征带的引物,共筛选出9对引物对后代进行真实性鉴定。把父母本和后代一起进行PCR扩增、电泳,根据后代中是否能扩增出父本特征带而进行鉴定。结果表明,37个杂交后代中有32个后代具有父本特征带,被鉴定为真杂种。同时,从杂种扩增的谱带来看,条带清晰、稳定且后代出现了丰富的变异,主要表现为新谱带的出现或某些谱带的消失。据此认为,SRAP标记可以有效地对结缕草属杂交后代进行真实性鉴定和遗传分析,为杂交育种提供了理论依据。