基于灰度和梯度不变矩的人脸识别

提出一种新的人脸识别方法———灰度-梯度共轭不变矩法,克服了传统方法的缺点,在不增加复杂度的前提下,将灰度信息和梯度信息有效地结合起来,并将其共轭分布反映的纹理特征作为对灰度特征的补充参与特征分类.同时,对经典不变矩组进行分析并改进,消除了离散状态下比例因子的影响.以BP网络作为分类器进行试验,结果表明,与传统方法相比,该方法具有更高的识别率.     

A QUANTIFIER-ELIMINATION BASED HEURISTIC FOR AUTOMATICALLY GENERATING INDUCTIVE ASSERTIONS FOR PROGRAMS

A method using quantifier-elimination is proposed for automatically generating programinvariants/inductive assertions.Given a program,inductive assertions,hypothesized as parameterizedformulas in a theory,are associated with program locations.Parameters in inductive assertions arediscovered by generating constraints on parameters by ensuring that an inductive assertion is indeedpreserved by all execution paths leading to the associated location of the program.The method can beused to discover loop invariants-properties of variables that remain invariant at the entry of a loop.Theparameterized formula can be successively refined by considering execution paths one by one;heuristicscan be developed for determining the order in which the paths are considered.Initialization of programvariables as well as the precondition and postcondition,if available,can also be used to further refinethe hypothesized invariant.The method does not depend on the availability of the precondition andpostcondition of a program.Constraints on parameters generated in this way are solved for possiblevalues of parameters.If no solution is possible,this means that an invariant of the hypothesizedform is not likely to exist for the loop under the assumptions/approximations made to generate theassociated verification condition.Otherwise,if the parametric constraints are solvable,then undercertain conditions on methods for generating these constraints,the strongest possible invariant of thehypothesized form can be generated from most general solutions of the parametric constraints.Theapproach is illustrated using the logical languages of conjunction of polynomial equations as well asPresburger arithmetic for expressing assertions.     

LF拓扑空间的S-连通性

给出了LF拓扑空间中的S—连通性的概念，证明了（L^X，δ）不是S—连通空间的二个等价条件；讨论了S—连通集的性质；论证了S—连通性的半同胚性质及S—连通分支的概念和性质。     

A Project Risk Ranking Approach Based on Set Pair Analysis

Set Pair Analysis （SPA） is a new methodology to describe and process system uncertainty. It is different from stochastic or fuzzy methods in reasoning and operation, and it has been applied in many areas recently. In this paper, the application of SPA in risk ranking is presented, which includes review of risk ranking, introduction of Connecting Degree （CD） that is a key role in SPA., Arithmetic and Tendency Grade （TG） of CDs, and a risk ranking approach proposed. Finally a case analysis is presented to illustrate the reasonability of this approach. It is found that this approach is very convenient to operate, while the ranking result is more comprehensible.     

基于粗糙集理论的知识库分解与规则获取

粗糙集理论可以通过对数据的约简从例子中学习,获取决策规则。但是,当知识库规模较大、条件属性个数较多时,存在提取规则速度慢、规则长度长等缺点。本文介绍了粗糙集理论的基本概念,提出了一种基于粗糙集的知识库分解算法。首先引入决策属性支持度的概念,在此基础上定义了一个属性选择量度,选择最佳目标属性对知识库进行分解,直到所有对象都被精确分类,从而得到具有一定支持度的规则集,得到的规则长度短,提取速度快。通过对一个简单实例的分析,证明了该算法的可行性。     

19GeV/c能量下质子-铍碰撞中领头奇异超子Λ的产生

利用在美国布鲁克海汶国家实验室的环型加速器(Brookhaven/AGS)上进行的19.6 GeV质子轰击固定铍核的实验数据,通过质子与π介子对的搭配,并用一种新的循环迭代的混合事件方法,从复杂的背景中,提炼出共振态领头粒子Λ的产生.     

核心元素的特征及其在决策树生成中的应用

利用粗糙集约简的辨识矩阵，给出属性是核心元素的充要条件为该属性是辨识矩阵中单元素集．利用上近似与下近似给出了判断是否核心元素的两个充分条件，指出利用核心元素简化决策树比用熵函数ID3方法简化的决策树更简单．     

关于量子力学中效应代数上的收敛理论及不变量

本文介绍了量子力学中效应代数的研究进展。指出利用无穷矩阵理论研究其上的收敛理论和不变量，对建立量子力学的数学基础有重要意义。     

基于粗集理论的一种规则提取方法

介绍了粗集理论的基本概念,通过对现有基于粗集的最小规则提取算法的分析,发现其缺陷,提出了一种改进的基于粗糙集的规则提取方法,并用UCI数据库中的9个标准数据集从规则集的规则数目、规则集的平均规则长度、规则集的平均规则支持、规则集的预测精度等4个指标对改进的算法进行了测试和对比分析,实验表明了该算法的有效性.     

一种改进的Rough集属性约简启发式遗传算法

属性约简是知识发现中的关键问题之一 .为了能够有效地获取决策表中属性最小相对约简 ,提出了一种在优化初始群体基础上提高算法性能的启发式遗传算法 .首先 ,通过构造一个新的算子 ,将信息论角度定义的属性重要性度量作为启发式信息 ,来描述所选择的属性子集对论域中确定分类子集的影响 ;接着 ,以此为基础并结合遗传算法 ,选择一些经过优化的染色体作为初始群体 ,在加强局部搜索能力的同时保持了该算法全局寻优的特性 .最后 ,从理论上对算法做了分析 ,证明了新算子所选择的属性子集对原有属性分类能力保持不变 .试验分析表明 ,该算法能有效地对决策表属性进行约简