According to this process , the key technology which including the characteristic measurement of surface , the pretreatment of data getting by line scan , the character expression and fitting of curve and surface , the characteristic identify of surface and translation from the model of surface to cad 对实物反求过程中的各关键技术作了研究，包括：曲面的特征测量、逐行扫描测量数据的预处理、曲线曲面的特征表示及其拟合、曲面的特征识别以及曲面模型向cad模型的转化。
Due to the unique , stable and live physiological properties of the iris and non - invasive to users , most stable and reliable for identification in practical applications of iris - based system , iris recognition is receiving extensive attention and becoming an active topic in biometrics . as an application - oriented research project , iris recognition integrates mathematic , computer science , optics , electronics and physiology etc . based on wavelet transform , aimed to improve the recognition performance , centered at wavelet - based iris feature representation and pattern classification , we review and explore the iris sequence image quality assessment , iris image pre - processing , iris recognition performance evaluation and several other linked topics . we mainly investigate on the principles and application methodology of wavelet transform for iris feature representation and iris pattern classification methods 以小波变换技术为基础,结合图像处理和模式识别方法,设计并开发了虹膜图像采集装置,建立了虹膜识别算法测试实验平台;重点研究了虹膜识别中的小波变换的应用基础理论与关键实现技术;提出了基于小波局部模极大值的虹膜特征表示及其多重匹配识别、基于小波多尺度信息的一维和二维虹膜纹理特征量化表示、基于小波过零点技术的虹膜特征表示及其规范化的部分hausdorff距离匹配识别,这三类方法能够有效地提取虹膜特征;基于自建的算法测试平台,对上述三类方法和其他三种国内外比较有影响的基于小波变换的虹膜识别方法进行了定量的性能比较和评价,通过实验数据分析得到了有意义的结论;最后指出了小波变换技术在虹膜识别领域的研究重点与发展方向。
This paper has a further study on the key technology called validity maintenance mechanism in semantic feature modeling in order to maintain the intent of designers and satisfy them . the method combines geometry model and semantic feature modeling , setting up new feature representation 、 cell element naming and identification method . in the meantime , a history - independence interactive feature boundary re - evaluation algorithm is presented based on feature editing . moreover , the validity recovery mechanism after invalid feature operation can maintain the original feature model . the above ideas have been realized on hust - caid ( computer aided industry design system developed by harbin university of science and technology ) preliminarily 为了能够正确地维护设计者的设计意图和满足用户的需求，本文在原有造型的基础上对语义特征造型中的关键技术即有效性维护机制进行深入的研究，将几何模型，语义特征模型结合起来统一进行研究，建立了语义特征造型在产品模型设计过程中特征表示方法的新理论、细胞元素命名方法和辨别机制，同时对系统原有特征编辑过程进行了研究，提出了独立于历程树的特征边界重构算法以及模型操作无效时的有效性恢复方法。
This provide a basis for further application such as feature editing and so on . unlike traditional method which modifies b - rep or csg to adapt feature - based modeling , the feature - based represent model for the product which based on identifying system is presented . the represent method need not change the b - rep of solid model and can map feature - based model to b - rep - based model 不同于通常的对边界表示法或csg法进行改造的模型表示方法，提出了一种基于标识子系统的特征表示模型，一个零件的特征表示模型和其低层的几何实体模型通过标识子系统进行映射，从而不必修改现有的边界表示模型而能满足特征造型的需要并能充分利用已有的实体造型功能。
Such integra ting feature vector is used for building k dim e nsion gaussian m odel , whose param e ters are estim ated by an expectation - m a xi m i zation ( em ) algorithm , and then the resulting block - cluster m e mberships provide a segm entation of th e im age . after segm ented , a m e thod of param e ter - trimm e d average for describing re gion is proposed , of which the param e ter is decided by area and position of region dire ctly . the sim ilarity m easure between two im ages is defined by integrating properties of all regions in the im age 文中先将图像分成4 4小块，各块的颜色、纹理、位置特征构成8维的特征空间；在该空间中对得到的8维特征矢量建立一个k维高斯模型，应用期望最大em算法估计模型参数，产生的块特征-聚类隶属度函数实现对图像的分割；为减小分割算法不确定性对检索效果的不良影响，对得到的区域采用参数均衡平均特征表示，其中参数的确定直接与区域的面积、位置有关。
Second , we analyze construction of streaming media and limited - relation inner streaming media data and its time - space feature expressing . at last , we analyze the system synchronizations and give the corresponding synchronize arithmetic and control strategy to inner streaming and between streaming and the groups . and we deduce the minimum of memory for every receiver and the starting time offset of the first ldu of every code streaming playing in every receiver and the concrete method for judgment whether realize the group synchronization 论文首先介绍了流媒体技术与流媒体同步目前的发展状况以及流媒体传输协议，接着分析了流媒体数据内部约束关系及流媒体数据的构成，探讨了流媒体数据时域特征表示，然后对该多对多流媒体系统的流内、流间及组同步问题分别进行分析研究，推导出接收端各信宿的缓存器所要具有的最小容量、各码流的第1个ldu在接收端各信宿起始播放时刻的偏移量以及组同步与否的具体判断方法，对多对多流媒体系统的流内、流间及组同步分别设计出各自的同步算法及控制策略。
This article discussed the chinese word slice , character extraction , character expression and character matching methods , and established the chinese text classification and clustering algorithms based on neural network . in the design of chinese text mining based on web , the paper analyzed and researched the expression of web page information , structure feature , web page control symbol and html control symbol , and built the extraction flow of web page information , then gave two concrete application of chinese text mining based on web through combining with practical problems 讨论了文本分类中的中文词切分、特征提取、特征表示、特征匹配方法，建立了基于神经网络的中文文本分类、聚类算法，在web中文文本信息挖掘的设计中，对网页信息的表示、结构特点、网页控制符、 html控制符号处理进行了详细分析与研究，构建了网页信息提取流程，并结合实际问题，给出了web环境下中文文本信息挖掘的两个具体应用。
