形状距离学习是形状匹配框架中引入的后处理步骤,能够有效改善逐对计算得到的形状间距离.利用期望首达时间分析形状间相似度可能导致距离更新不准确,针对这一问题提出了一种基于广义期望首达时间(Generalized mean firstpassage time,GM...形状距离学习是形状匹配框架中引入的后处理步骤,能够有效改善逐对计算得到的形状间距离.利用期望首达时间分析形状间相似度可能导致距离更新不准确,针对这一问题提出了一种基于广义期望首达时间(Generalized mean firstpassage time,GMFPT)的形状距离学习方法.将形状样本集合视作状态空间,广义期望首达时间表示质点由一个状态转移至指定状态集合所需的平均时间步长,本文将其视作更新后的形状间距离.通过引入广义期望首达时间,形状距离学习方法能够有效地分析上下文相关的形状相似度,显式地挖掘样本空间流形中的最短路径,并消除冗余上下文形状信息的影响.将所提出的方法应用到不同形状数据集中进行仿真实验,本文方法比其他方法能够得到更准确的形状检索结果.展开更多
A new method that uses a modified ordered subsets （MOS） algorithm to improve the convergence rate of space-alternating generalized expectation-maximization （SAGE） algorithm for positron emission tomography （PET）...A new method that uses a modified ordered subsets （MOS） algorithm to improve the convergence rate of space-alternating generalized expectation-maximization （SAGE） algorithm for positron emission tomography （PET） image reconstruction is proposed.In the MOS-SAGE algorithm,the number of projections and the access order of the subsets are modified in order to improve the quality of the reconstructed images and accelerate the convergence speed.The number of projections in a subset increases as follows：2,4,8,16,32 and 64.This sequence means that the high frequency component is recovered first and the low frequency component is recovered in the succeeding iteration steps.In addition,the neighboring subsets are separated as much as possible so that the correlation of projections can be decreased and the convergences can be speeded up.The application of the proposed method to simulated and real images shows that the MOS-SAGE algorithm has better performance than the SAGE algorithm and the OSEM algorithm in convergence and image quality.展开更多
美国资本市场历史数据显示股票收益远远高于债券收益,不能用基于消费的资产定价模型解释,Mehra and Prescott提出了所谓的股权溢价之谜,而Campbell and Cochrane的习惯形成模型在用HJ方差界检验股权溢价之谜时受到局限。对Epstein and Z...美国资本市场历史数据显示股票收益远远高于债券收益,不能用基于消费的资产定价模型解释,Mehra and Prescott提出了所谓的股权溢价之谜,而Campbell and Cochrane的习惯形成模型在用HJ方差界检验股权溢价之谜时受到局限。对Epstein and Zin(1991)的广义期望效用模型进行修正,使用随机贴现因子的HJ方差界来检验中国股票市场是否存在股权溢价之谜,并比较CRRA模型、Epstein and Zin的模型及修正模型的定价能力。实证分析发现:(1)修正后的模型要求消费者的相对风险厌恶系数为2左右就可以解释中国的高股权溢价现象,而且中国股票市场不存在股权溢价之谜,也不存在无风险利率之谜;(2)同CRRA模型、Epstein and Zin的模型相比,修正后的模型具有更强的定价能力。展开更多
The one-block version of ordered subsets (OS) techniques was used to accelerate the convergent rate of the space-alternating generalized expectation-maximization (SAGE) algorithm. The new row-action SAGE (RA-SAGE) alg...The one-block version of ordered subsets (OS) techniques was used to accelerate the convergent rate of the space-alternating generalized expectation-maximization (SAGE) algorithm. The new row-action SAGE (RA-SAGE) algorithm processed projections in sequentially orthogonal order which reduced the dependency among the projections and speeds up the convergences. Additionally, the over-relaxation parameter in the direction defined by the RA-SAGE algorithm was also applied to obtain fast convergence to a globally maximum likelihood (ML) solution. In experiments, the RA-SAGE algorithm and the classical SAGE algorithm were compared in the application to positron emission tomography (PET) image reconstruction. Simulation results showed that RA-SAGE had better performance than SAGE in both convergence and image quality.展开更多
文摘形状距离学习是形状匹配框架中引入的后处理步骤,能够有效改善逐对计算得到的形状间距离.利用期望首达时间分析形状间相似度可能导致距离更新不准确,针对这一问题提出了一种基于广义期望首达时间(Generalized mean firstpassage time,GMFPT)的形状距离学习方法.将形状样本集合视作状态空间,广义期望首达时间表示质点由一个状态转移至指定状态集合所需的平均时间步长,本文将其视作更新后的形状间距离.通过引入广义期望首达时间,形状距离学习方法能够有效地分析上下文相关的形状相似度,显式地挖掘样本空间流形中的最短路径,并消除冗余上下文形状信息的影响.将所提出的方法应用到不同形状数据集中进行仿真实验,本文方法比其他方法能够得到更准确的形状检索结果.
基金The National Basic Research Program of China (973Program) (No.2003CB716102).
文摘A new method that uses a modified ordered subsets （MOS） algorithm to improve the convergence rate of space-alternating generalized expectation-maximization （SAGE） algorithm for positron emission tomography （PET） image reconstruction is proposed.In the MOS-SAGE algorithm,the number of projections and the access order of the subsets are modified in order to improve the quality of the reconstructed images and accelerate the convergence speed.The number of projections in a subset increases as follows：2,4,8,16,32 and 64.This sequence means that the high frequency component is recovered first and the low frequency component is recovered in the succeeding iteration steps.In addition,the neighboring subsets are separated as much as possible so that the correlation of projections can be decreased and the convergences can be speeded up.The application of the proposed method to simulated and real images shows that the MOS-SAGE algorithm has better performance than the SAGE algorithm and the OSEM algorithm in convergence and image quality.
文摘美国资本市场历史数据显示股票收益远远高于债券收益,不能用基于消费的资产定价模型解释,Mehra and Prescott提出了所谓的股权溢价之谜,而Campbell and Cochrane的习惯形成模型在用HJ方差界检验股权溢价之谜时受到局限。对Epstein and Zin(1991)的广义期望效用模型进行修正,使用随机贴现因子的HJ方差界来检验中国股票市场是否存在股权溢价之谜,并比较CRRA模型、Epstein and Zin的模型及修正模型的定价能力。实证分析发现:(1)修正后的模型要求消费者的相对风险厌恶系数为2左右就可以解释中国的高股权溢价现象,而且中国股票市场不存在股权溢价之谜,也不存在无风险利率之谜;(2)同CRRA模型、Epstein and Zin的模型相比,修正后的模型具有更强的定价能力。
文摘The one-block version of ordered subsets (OS) techniques was used to accelerate the convergent rate of the space-alternating generalized expectation-maximization (SAGE) algorithm. The new row-action SAGE (RA-SAGE) algorithm processed projections in sequentially orthogonal order which reduced the dependency among the projections and speeds up the convergences. Additionally, the over-relaxation parameter in the direction defined by the RA-SAGE algorithm was also applied to obtain fast convergence to a globally maximum likelihood (ML) solution. In experiments, the RA-SAGE algorithm and the classical SAGE algorithm were compared in the application to positron emission tomography (PET) image reconstruction. Simulation results showed that RA-SAGE had better performance than SAGE in both convergence and image quality.
