城市功能区识别可为城市建设决策提供技术支撑。本研究提出了面向城市功能区识别的多源场景特征Transformer融合方法。利用路网构建交通分析区(Traffic Analysis Zone,TAZ),采用Delaunay三角网创建POI(Point of Interest)数据的图结构,...城市功能区识别可为城市建设决策提供技术支撑。本研究提出了面向城市功能区识别的多源场景特征Transformer融合方法。利用路网构建交通分析区(Traffic Analysis Zone,TAZ),采用Delaunay三角网创建POI(Point of Interest)数据的图结构,通过TAZ获得遥感数据的影像对象;利用图卷积网络提取POI图结构的社会场景特征,由ResNet-50编码遥感数据的自然场景特征;基于Transformer的多头注意力机制融合多维特征,依托SoftMax实现功能区识别。以沈阳市主城区为例,以2021年的OpenStreetMap、POI和遥感数据为数据源。该方法的总体精度和Kappa系数为82.2%和70%,Kappa系数较单一数据方法和其他融合方法至少提高18%和9%。本研究采用Transformer融合社会场景特征和自然场景特征,解决了多源数据难以集成表达的问题,为城市功能区识别提供了新的技术路径。展开更多
目前自闭症功能磁共振(functional magnetic resonance imaging,fMRI)图像分类模型在跨多个机构的数据集下分类精度较低,难以应用到自闭症的诊断工作中。为此,本文提出了一种基于Transformer的自闭症分类模型(autism spectrum disorder ...目前自闭症功能磁共振(functional magnetic resonance imaging,fMRI)图像分类模型在跨多个机构的数据集下分类精度较低,难以应用到自闭症的诊断工作中。为此,本文提出了一种基于Transformer的自闭症分类模型(autism spectrum disorder classification model based on Transformer,TransASD)。首先采用脑图谱模板提取fMRI数据中的时间序列输入Transformer模型,并引入一种重叠窗口注意力机制,能够更好地捕捉异构数据的局部与全局特征。其次,提出了一个跨窗口正则化方法作为额外的损失项,使模型可以更加准确地聚焦于重要的特征。本文使用该模型在公开的自闭症数据集ABIDE上进行实验,在10折交叉验证法下得到了71.44%的准确率,该模型对比其他先进算法模型取得了更好的分类效果。展开更多
Let and denote respectively the functionswhere λ≥1, The author discusses the similarity transformation of the regularizing functionals of these functions and the similar property of their Fourier transformation.
针对多变量时序(Multivariate Time Series,MTS)分类中长序列数据难以捕捉时序特征的问题,提出一种基于双向稀疏Transformer的时序分类模型BST(Bidirectional Sparse Transformer),提高了MTS分类任务的准确度.BST模型使用Transformer框...针对多变量时序(Multivariate Time Series,MTS)分类中长序列数据难以捕捉时序特征的问题,提出一种基于双向稀疏Transformer的时序分类模型BST(Bidirectional Sparse Transformer),提高了MTS分类任务的准确度.BST模型使用Transformer框架,构建了一种基于活跃度得分的双向稀疏注意力机制.基于KL散度构建活跃度评价函数,并将评价函数的非对称问题转变为对称权重问题.据此,对原有查询矩阵、键值矩阵进行双向稀疏化,从而降低原Transformer模型中自注意力机制运算的时间复杂度.实验结果显示,BST模型在9个长序列数据集上取得最高平均排名,在临界差异图中领先第2名35.7%,对于具有强时序性的乙醇浓度数据集(Ethanol Concentration,EC),分类准确率提高30.9%.展开更多
由于水下环境的多样性和光在水中受到的散射及选择性吸收作用,采集到的水下图像通常会产生严重的质量退化问题,如颜色偏差、清晰度低和亮度低等,为解决以上问题,本文提出了一种基于Transformer和生成对抗网络的水下图像增强算法。以生...由于水下环境的多样性和光在水中受到的散射及选择性吸收作用,采集到的水下图像通常会产生严重的质量退化问题,如颜色偏差、清晰度低和亮度低等,为解决以上问题,本文提出了一种基于Transformer和生成对抗网络的水下图像增强算法。以生成对抗网络为基础架构,结合编码解码结构、基于空间自注意力机制的全局特征建模Transformer模块和通道级多尺度特征融合Transformer模块构建了TGAN(generative adversarial network with transformer)网络增强模型,重点关注水下图像衰减更严重的颜色通道和空间区域,有效增强了图像细节并解决了颜色偏差问题。此外,设计了一种结合RGB和LAB颜色空间的多项损失函数,约束网络增强模型的对抗训练。实验结果表明,与CLAHE(contrast limited adaptive histogram equalization)、UDCP(underwater dark channel prior)、UWCNN(underwater based on convolutional neural network)、FUnIE-GAN(fast underwater image enhancement for improved visual perception)等典型水下图像增强算法相比,所提算法增强后的水下图像在清晰度、细节纹理和色彩表现等方面都有所提升,客观评价指标如峰值信噪比、结构相似性和水下图像质量度量的平均值分别提升了5.8%、1.8%和3.6%,有效地提升了水下图像的视觉感知效果。展开更多
In China, most directions of river flowing are from west to east. During historic period, since the water traffic played an important role, it was very important to form a cross-horizontal net of water carriage route....In China, most directions of river flowing are from west to east. During historic period, since the water traffic played an important role, it was very important to form a cross-horizontal net of water carriage route. Canals should be dug so as to make up the lack of north-south river. Tongji Grand Canal, dug in the first year of Daye (605 A.D.) in the Sui Dynasty, was the important component of north-south system of Grand Canals in China. It promoted economic and social development of the Sui, Tang and Song dynasties (605 A.D.-1279 A.D.). As Tongji Canal (i.e. Tongji Grand Canal) flowed across the Huaibei Plain, which is aggraded by abundant mud and sand deposit resulted from the Huanghe (Yellow) River flooding, many traces (such as old channel) and human culture heritages were buried under mud-sand. Tongji Canal was silted up, and disappeared in the Jin Dynasty (1115 A.D.-1234 A.D.). From then on, there were many different stories about the flowing route of the canal in historical literature. Based on space-bone and air-bone remote sensing imagery, we attempt to search the old channel of Tongji Canal, and supplement historical record. The paper discusses transformation process of Tongji Canal’s traffic function, and resumptively summarizes the reasons of the transformation, which results from synthetic function of physical geographical, political, economic, and social conditions.展开更多
Reliability allocation of computerized numerical controlled（CNC）lathes is very important in industry.Traditional allocation methods only focus on high-failure rate components rather than moderate failure rate compon...Reliability allocation of computerized numerical controlled（CNC）lathes is very important in industry.Traditional allocation methods only focus on high-failure rate components rather than moderate failure rate components,which is not applicable in some conditions.Aiming at solving the problem of CNC lathes reliability allocating,a comprehensive reliability allocation method based on cubic transformed functions of failure modes and effects analysis（FMEA）is presented.Firstly,conventional reliability allocation methods are introduced.Then the limitations of direct combination of comprehensive allocation method with the exponential transformed FMEA method are investigated.Subsequently,a cubic transformed function is established in order to overcome these limitations.Properties of the new transformed functions are discussed by considering the failure severity and the failure occurrence.Designers can choose appropriate transform amplitudes according to their requirements.Finally,a CNC lathe and a spindle system are used as an example to verify the new allocation method.Seven criteria are considered to compare the results of the new method with traditional methods.The allocation results indicate that the new method is more flexible than traditional methods.By employing the new cubic transformed function,the method covers a wider range of problems in CNC reliability allocation without losing the advantages of traditional methods.展开更多
In this paper, an extended functional transformation is given to solve some nonlinear evolution equations. This function, in fact, is a solution of the famous KdV equation, so this transformation gives a transformatio...In this paper, an extended functional transformation is given to solve some nonlinear evolution equations. This function, in fact, is a solution of the famous KdV equation, so this transformation gives a transformation between KdV equation and other soliton equations. Then many new exact solutions can be given by virtue of the solutions of KdV equation.展开更多
The theory of elastic wave scattering is a fundamental concept in the study of elastic dynamics and wave motion,and the wave function expansion technique has been widely used in many subjects.To supply the essential t...The theory of elastic wave scattering is a fundamental concept in the study of elastic dynamics and wave motion,and the wave function expansion technique has been widely used in many subjects.To supply the essential tools for solving wave scattering problems induced by an eccentric source or multi-sources as well as multi-scatters,a whole-space transform formula of cylindrical wave functions is presented and its applicability to some simple cases is demonstrated in this study.The transforms of wave functions in cylindrical coordinates can be classifi ed into two basic types: interior transform and exterior transform,and the existing Graf’s addition theorem is only suitable for the former.By performing a new replacement between the two coordinates,the exterior transform formula is fi rst deduced.It is then combined with Graf’s addition theorem to establish a whole-space transform formula.By using the whole-space transform formula,the scattering solutions by the sources outside and inside a cylindrical cavity are constructed as examples of its application.The effectiveness and advantages of the whole-space transform formula is illustrated by comparison with the approximate model based on a large cycle method.The whole-space transform formula presented herein can be used to perform the transform between two different cylindrical coordinates in the whole space.In addition,its concept and principle are universal and can be further extended to establish the coordinate transform formula of wave functions in other coordinate systems.展开更多
The probability distributions of wave characteristics from three groups of sampled ocean data with different significant wave heights have been analyzed using two transformation functions estimated by non-parametric a...The probability distributions of wave characteristics from three groups of sampled ocean data with different significant wave heights have been analyzed using two transformation functions estimated by non-parametric and parametric methods. The marginal wave characteristic distribution and the joint density of wave properties have been calculated using the two transformations, with the results and accuracy of both transformations presented here. The two transformations deviate slightly between each other for the calculation of the crest and trough height marginal wave distributions, as well as the joint densities of wave amplitude with other wave properties. The transformation methods for the calculation of the wave crest and trough height distributions are shown to provide good agreement with real ocean data. Our work will help in the determination of the most appropriate transformation procedure for the prediction of extreme values.展开更多
The Auto-Transformer Rectifier Unit（ATRU） is one preferred solution for high-power AC/DC power conversion in aircraft. This is mainly due to its simple structure, high reliability and reduced k VA ratings. Indeed, t...The Auto-Transformer Rectifier Unit（ATRU） is one preferred solution for high-power AC/DC power conversion in aircraft. This is mainly due to its simple structure, high reliability and reduced k VA ratings. Indeed, the ATRU has become a preferred AC/DC solution to supply power to the electric environment control system on-board future aircraft. In this paper, a general modelling method for ATRUs is introduced. The developed model is based on the fact that the DC voltage and current are strongly related to the voltage and current vectors at the AC terminals of ATRUs. In this paper, we carry on our research in modelling symmetric 18-pulse ATRUs and develop a generic modelling technique. The developed generic model can study not only symmetric but also asymmetric ATRUs. An 18-pulse asymmetric ATRU is used to demonstrate the accuracy and efficiency of the developed model by comparing with corresponding detailed switching SABER models provided by our industrial partner. The functional models also allow accelerated and accurate simulations and thus enable whole-scale more-electric aircraft electrical power system studies in the future.展开更多
In order to provide larger capacity of the hidden secret data while maintaining a good visual quality of stego-image, in accordance with the visual property that human eyes are less sensitive to strong texture, a nove...In order to provide larger capacity of the hidden secret data while maintaining a good visual quality of stego-image, in accordance with the visual property that human eyes are less sensitive to strong texture, a novel steganographic method based on wavelet and modulus function is presented. First, an image is divided into blocks of prescribed size, and every block is decomposed into one-level wavelet. Then, the capacity of the hidden secret data is decided with the number of wavelet coefficients of larger magnitude. Finally, secret information is embedded by steganography based on modulus function. From the experimental results, the proposed method hides much more information and maintains a good visual quality of stego-image. Besides, the embedded data can be extracted from the stego-image without referencing the original image.展开更多
文摘城市功能区识别可为城市建设决策提供技术支撑。本研究提出了面向城市功能区识别的多源场景特征Transformer融合方法。利用路网构建交通分析区(Traffic Analysis Zone,TAZ),采用Delaunay三角网创建POI(Point of Interest)数据的图结构,通过TAZ获得遥感数据的影像对象;利用图卷积网络提取POI图结构的社会场景特征,由ResNet-50编码遥感数据的自然场景特征;基于Transformer的多头注意力机制融合多维特征,依托SoftMax实现功能区识别。以沈阳市主城区为例,以2021年的OpenStreetMap、POI和遥感数据为数据源。该方法的总体精度和Kappa系数为82.2%和70%,Kappa系数较单一数据方法和其他融合方法至少提高18%和9%。本研究采用Transformer融合社会场景特征和自然场景特征,解决了多源数据难以集成表达的问题,为城市功能区识别提供了新的技术路径。
文摘目前自闭症功能磁共振(functional magnetic resonance imaging,fMRI)图像分类模型在跨多个机构的数据集下分类精度较低,难以应用到自闭症的诊断工作中。为此,本文提出了一种基于Transformer的自闭症分类模型(autism spectrum disorder classification model based on Transformer,TransASD)。首先采用脑图谱模板提取fMRI数据中的时间序列输入Transformer模型,并引入一种重叠窗口注意力机制,能够更好地捕捉异构数据的局部与全局特征。其次,提出了一个跨窗口正则化方法作为额外的损失项,使模型可以更加准确地聚焦于重要的特征。本文使用该模型在公开的自闭症数据集ABIDE上进行实验,在10折交叉验证法下得到了71.44%的准确率,该模型对比其他先进算法模型取得了更好的分类效果。
文摘Let and denote respectively the functionswhere λ≥1, The author discusses the similarity transformation of the regularizing functionals of these functions and the similar property of their Fourier transformation.
文摘由于水下环境的多样性和光在水中受到的散射及选择性吸收作用,采集到的水下图像通常会产生严重的质量退化问题,如颜色偏差、清晰度低和亮度低等,为解决以上问题,本文提出了一种基于Transformer和生成对抗网络的水下图像增强算法。以生成对抗网络为基础架构,结合编码解码结构、基于空间自注意力机制的全局特征建模Transformer模块和通道级多尺度特征融合Transformer模块构建了TGAN(generative adversarial network with transformer)网络增强模型,重点关注水下图像衰减更严重的颜色通道和空间区域,有效增强了图像细节并解决了颜色偏差问题。此外,设计了一种结合RGB和LAB颜色空间的多项损失函数,约束网络增强模型的对抗训练。实验结果表明,与CLAHE(contrast limited adaptive histogram equalization)、UDCP(underwater dark channel prior)、UWCNN(underwater based on convolutional neural network)、FUnIE-GAN(fast underwater image enhancement for improved visual perception)等典型水下图像增强算法相比,所提算法增强后的水下图像在清晰度、细节纹理和色彩表现等方面都有所提升,客观评价指标如峰值信噪比、结构相似性和水下图像质量度量的平均值分别提升了5.8%、1.8%和3.6%,有效地提升了水下图像的视觉感知效果。
基金Under the auspices of the National Natural Science Foundation of China (No.40571162), the Natural Science Foun-dation of Anhui Province (No.050450401), Natural Science Foundation of Anhui Provincal Education Department (No.2004kj65zd), Postdoctoral Foundation of University of Science and Technology of China
文摘In China, most directions of river flowing are from west to east. During historic period, since the water traffic played an important role, it was very important to form a cross-horizontal net of water carriage route. Canals should be dug so as to make up the lack of north-south river. Tongji Grand Canal, dug in the first year of Daye (605 A.D.) in the Sui Dynasty, was the important component of north-south system of Grand Canals in China. It promoted economic and social development of the Sui, Tang and Song dynasties (605 A.D.-1279 A.D.). As Tongji Canal (i.e. Tongji Grand Canal) flowed across the Huaibei Plain, which is aggraded by abundant mud and sand deposit resulted from the Huanghe (Yellow) River flooding, many traces (such as old channel) and human culture heritages were buried under mud-sand. Tongji Canal was silted up, and disappeared in the Jin Dynasty (1115 A.D.-1234 A.D.). From then on, there were many different stories about the flowing route of the canal in historical literature. Based on space-bone and air-bone remote sensing imagery, we attempt to search the old channel of Tongji Canal, and supplement historical record. The paper discusses transformation process of Tongji Canal’s traffic function, and resumptively summarizes the reasons of the transformation, which results from synthetic function of physical geographical, political, economic, and social conditions.
基金Supported by National Natural Science Foundation of China(Grant Nos.51135003,51205050,U1234208)Key National Science & Technology Special Project on"High-Grade CNC Machine Tools and Basic Manufacturing Equipments"(Grant No.2013ZX04011011)+1 种基金Research Fund for the Doctoral Program of Higher Education of China(Grant No.20110042120020)Fundamental Research Funds for the Central
文摘Reliability allocation of computerized numerical controlled（CNC）lathes is very important in industry.Traditional allocation methods only focus on high-failure rate components rather than moderate failure rate components,which is not applicable in some conditions.Aiming at solving the problem of CNC lathes reliability allocating,a comprehensive reliability allocation method based on cubic transformed functions of failure modes and effects analysis（FMEA）is presented.Firstly,conventional reliability allocation methods are introduced.Then the limitations of direct combination of comprehensive allocation method with the exponential transformed FMEA method are investigated.Subsequently,a cubic transformed function is established in order to overcome these limitations.Properties of the new transformed functions are discussed by considering the failure severity and the failure occurrence.Designers can choose appropriate transform amplitudes according to their requirements.Finally,a CNC lathe and a spindle system are used as an example to verify the new allocation method.Seven criteria are considered to compare the results of the new method with traditional methods.The allocation results indicate that the new method is more flexible than traditional methods.By employing the new cubic transformed function,the method covers a wider range of problems in CNC reliability allocation without losing the advantages of traditional methods.
文摘In this paper, an extended functional transformation is given to solve some nonlinear evolution equations. This function, in fact, is a solution of the famous KdV equation, so this transformation gives a transformation between KdV equation and other soliton equations. Then many new exact solutions can be given by virtue of the solutions of KdV equation.
基金the National Natural Science Foundation of China under Grand No.549974011
文摘The theory of elastic wave scattering is a fundamental concept in the study of elastic dynamics and wave motion,and the wave function expansion technique has been widely used in many subjects.To supply the essential tools for solving wave scattering problems induced by an eccentric source or multi-sources as well as multi-scatters,a whole-space transform formula of cylindrical wave functions is presented and its applicability to some simple cases is demonstrated in this study.The transforms of wave functions in cylindrical coordinates can be classifi ed into two basic types: interior transform and exterior transform,and the existing Graf’s addition theorem is only suitable for the former.By performing a new replacement between the two coordinates,the exterior transform formula is fi rst deduced.It is then combined with Graf’s addition theorem to establish a whole-space transform formula.By using the whole-space transform formula,the scattering solutions by the sources outside and inside a cylindrical cavity are constructed as examples of its application.The effectiveness and advantages of the whole-space transform formula is illustrated by comparison with the approximate model based on a large cycle method.The whole-space transform formula presented herein can be used to perform the transform between two different cylindrical coordinates in the whole space.In addition,its concept and principle are universal and can be further extended to establish the coordinate transform formula of wave functions in other coordinate systems.
基金Supported by the Marine Engineering Equipment Scientific Research Project of Ministry of Industry and Information Technology of PRCthe National Science and Technology Major Project of China(Grant No.2016ZX05057020)National Natural Science Foundation of China(Grant No.51809067)
文摘The probability distributions of wave characteristics from three groups of sampled ocean data with different significant wave heights have been analyzed using two transformation functions estimated by non-parametric and parametric methods. The marginal wave characteristic distribution and the joint density of wave properties have been calculated using the two transformations, with the results and accuracy of both transformations presented here. The two transformations deviate slightly between each other for the calculation of the crest and trough height marginal wave distributions, as well as the joint densities of wave amplitude with other wave properties. The transformation methods for the calculation of the wave crest and trough height distributions are shown to provide good agreement with real ocean data. Our work will help in the determination of the most appropriate transformation procedure for the prediction of extreme values.
文摘The Auto-Transformer Rectifier Unit（ATRU） is one preferred solution for high-power AC/DC power conversion in aircraft. This is mainly due to its simple structure, high reliability and reduced k VA ratings. Indeed, the ATRU has become a preferred AC/DC solution to supply power to the electric environment control system on-board future aircraft. In this paper, a general modelling method for ATRUs is introduced. The developed model is based on the fact that the DC voltage and current are strongly related to the voltage and current vectors at the AC terminals of ATRUs. In this paper, we carry on our research in modelling symmetric 18-pulse ATRUs and develop a generic modelling technique. The developed generic model can study not only symmetric but also asymmetric ATRUs. An 18-pulse asymmetric ATRU is used to demonstrate the accuracy and efficiency of the developed model by comparing with corresponding detailed switching SABER models provided by our industrial partner. The functional models also allow accelerated and accurate simulations and thus enable whole-scale more-electric aircraft electrical power system studies in the future.
基金the National Natural Science Foundation of China (50677014)Hunan Provincial Natural Science Foundation of China (06JJ50114).
文摘In order to provide larger capacity of the hidden secret data while maintaining a good visual quality of stego-image, in accordance with the visual property that human eyes are less sensitive to strong texture, a novel steganographic method based on wavelet and modulus function is presented. First, an image is divided into blocks of prescribed size, and every block is decomposed into one-level wavelet. Then, the capacity of the hidden secret data is decided with the number of wavelet coefficients of larger magnitude. Finally, secret information is embedded by steganography based on modulus function. From the experimental results, the proposed method hides much more information and maintains a good visual quality of stego-image. Besides, the embedded data can be extracted from the stego-image without referencing the original image.
