基于...

Our second kind of method is based on lineal transformation, including principle component analysis and discriminant analysis.

我们提出的第二类方法是基于线性变换的分类方法，包括基于主成分分析和辨别分析的方法。

A novel approach of binary component discriminant function is proposed by generalizing the original quadratic-form distance metric.

针对现有基于核函数的相关反馈方法处理样本不对称性时的缺陷，提出了基于超球面的支持向量分类方法。

The validity of the new numerical method is improved and the application is extended to dispersive media.

在此基础上，结合数字信号处理给出了基于复介电常数的〓TD-PF法和基于电导率的〓TD-CF法。

Then the similarity between them is calculated based on this distance matrix. To validate this retrieval approach, we test it on a database containing about four hundreds motion clips which can be divided into various kinds of human movements. The experimental results demonstrate the effectiveness of our approach.

实验结果表明，与现有大多数基于内容的三维人体运动数据检索方法相比，该方法能够取得更好的时间效率且不依赖于任何预先设定的参数，相对于现有基于关键帧的运动检索技术而言，其能够获得更好的检索效果。

Face detection system A new design of face detection system is proposed based upon color-based skin-like region segmentation and neural network-based face distinguishment.

人脸检测系统在对基于颜色的皮肤区域分割和基于神经网络的人脸判别方法进行研究的基础上，本文提出一种新的人脸检测系统设计。

So it is difficult to distinguish them using traditional neural network classification. A face distinguishment approach based on multi-stage neural networks is proposed.

为此，本文提出一种基于多级分类神经网络级联的人脸判别方法，通过把基于不同颜色分量的多级神经网络级联起来，并利用不同颜色分量由于来源、定义、算法的不同而体现出来的不同特性，以不同的颜色分量图象作为各级神经网络的分类特征，有效提高了人脸判别的正确率。

In this paper, we present a detection technology based on intrusion events in order to overcome the shortcomings of high false alert rate, serious redundant alerts and poor expression about the intrusion activities in today's intrusion detection systems. We describe the design ideas about the important methods used in this technology such as multi-sensor data fusion, raw alert distinguishment based on fuzzy rules and so on in detail.

在本文中，我们针对目前入侵检测系统普遍存在的误报率高、报警信息冗余现象严重和对入侵表述能力不强的缺陷，提出了一种基于入侵事件的检测分析技术，详细描述了该技术所采用主要方法的设计思想，如多传感器数据融合以及基于模糊规则的原始报警信息鉴别等。

Firstly, the M-ary modulated Min-Sum algorithm which based on Maximum Liklihood is provided, however, can not be applied to actual decoding due to its complexity. By applying the so called Generalized Distributive Law, we simplified the Min-Sum decoding for M-ary modulated LDPC codes based on Maximum Likelihood, and obtained what we called Enhanced Min-Sum algorithm.

首先基于最大似然的原理给出了多进制调制下的Min-Sum算法，然后利用广义分配率的原理对基于最大似然译码的多进制Min-Sum算法进行了简化，首次提出了多进制调制下LDPC码的改进Min-Sum算法。

The disyllabic pitch model is obtained with the help of cubic polynomial approximation and neural network technique.

双音节音高模型是韵律调节的基础；将Kohonen自组织神经网络技术应用到音高模型的分类，并取得了较好的分类效果；（4）基于正弦模型表示的汉语语音分析与合成：将正弦模型引入到汉语的语音分析合成中来，提出了汉语的基于正弦模型表达的分析合成算法，并对在汉语普通话语音合成中的应用，包括汉语四声的变调、时长音高的调整等进行了深入的研究；对于合成音节之间的拼接问题作了探讨，并提出了自己的观点；在国内外的文献中，还没有对于该方法用于汉语语音的分析合成的应用的报道，因此是本文的主要贡献之一。

We give a classifier design method based on sun-class division, bring forward a new divisibility rule, and can ascertain the number of sub-class by training.

给出了一种基于子类划分的分类器设计方法，提出了一个基于类内散布矩阵和类间散布矩阵的新的可分性准则，可以通过训练自动确定子类个数。

We have no common name for a mime of Sophron or Xenarchus and a Socratic Conversation; and we should still be without one even if the imitation in the two instances were in trimeters or elegiacs or some other kind of verse--though it is the way with people to tack on 'poet' to the name of a metre, and talk of elegiac-poets and epic-poets, thinking that they call them poets not by reason of the imitative nature of their work, but indiscriminately by reason of the metre they write in.

索夫农 、森那库斯和苏格拉底式的对话采用的模仿没有一个公共的名称；三音步诗、挽歌体或其他类型的诗的模仿也没有——人们把&诗人&这一名词和格律名称结合到一起，称之为挽歌体诗人或者史诗诗人，他们被称为诗人，似乎只是因为遵守格律写作，而非他们作品的模仿本质。

The relationship between communicative competence and grammar teaching should be that of the ends and the means.

交际能力和语法的关系应该是目标与途径的关系。