在算法上

This paper proposes and investigates fast variable size block motion estimation algorithm based on merge and split procedures based on H.

本文提出了一种基于分裂与合并的快速运动估计和帧间模式选择算法，该算法以8×8为初始块尺寸，并以8×8块进行运动搜索，在此基础上，根据块间运动向量的相关性，结合块的边缘特性完成运动估计和块模式选择。

Throught adaptive algorithm of network, the system constantly revise right values distribute so as to meet goal. Distribute the heuristic knowledge and the experience knowledge to network interconnection and right value.?The reasoning machine carries on reasoning according to the knowledge that involve in knowledge base and fact that users offer.?Through front computation of nerve network, the output port obtains the output vector.?The explanation module of ES is explained to user.At last, this paper has realized grab hydraulic system based on ANNES intelligent diagnosis, it has proven this fault diagnosis method feasibility.

ANNES的工作过程是：①首先建立知识库；②学习系统获取知识时，它只要求专家提供范例及相应的解，通过神经网络学习算法对样本进行学习，经过网络内部自适应算法不断修改权值分布以达到要求，把专家求解实际问题的启发式知识和经验分布到网络的互连及权值分布上；③推理机根据知识库中的知识和用户提供的事实进行推理；④通过神经网络的前向计算，在输出端获得输出矢量：⑤经专家系

The results show that the fuzzy control algorithm is adaptive for the changes of mechanism friction and oil viscidity.

为修正模糊规则和验证控制算法的效果，在建立的换挡模型上使用模糊控制算法进行仿真。

It also improves the AES algorithm in order to fit the different output data rate of vocoder. This paper analyzed the characteristics of the voice communication system firstly.

本文首先对语音通信系统的特点进行了分析，在此基础上提出了如何选择对称密钥密码算法和公开密钥密码算法。

In this thesis, LPC-10 and MELP algorithms etc. are deeply studied. Then a interpolation vocoder using relativity between frame is proposed in this thesis, which based on the soul and some modifications of these algorithms.

本文对美国军队联邦标准LPC-10、MELP等算法进行了深入的研究，吸取了各算法的精髓并部分改进，在此基础上利用帧间相关性实现一种帧间插值声码器。

Based on the analysis to the particularity of the cruise missile attacking, a complete recognition method based on the knowledge is developed. A new fractal-based infrared image feature extraction method is presented and the detailed theoretic analysis and implement procedure of this algorithm is submitted and tested in some experiments. With the limitation of the line detection based on traditional Hough transform that the information of the length and the end points of the line is unavailable, a new algorithm which makes use of the accessional strategy based on precognition information is put forward to meet the demand for more information of the line, simulation results show this method is effective. Finally, the whole process of airport target recognition is presented and the result images are also given.

使用分形方法提取目标的特征，在知识指导下，提出了一种基于目标特征模型的降维的形态学分形维数计算方法，对传统分形方法进行了改进，从理论上推证了算法的合理性，并对算法进行了仿真分析；针对传统Hough变换无法获得线段端点和长度信息的局限性，提出了一种基于目标特征先验知识的Hough变换融合策略，通过引入目标先验知识，可以有效地获得直线信息；对信息多而复杂的机场目标采用基于知识的目标识别方法，使用置信度模型实现不确定推理，对目标进行识别判断，将知识贯穿于整个识别过程中，对目标进行了有效地识别。

By analyzing the results of LSNR measurements for samples with different components, and comparing with the media of Aprilis Corporation, we give the suggestion to optimize the material components in the future. In addition, based on the physical mechanism of intra-page crosstalk noise, we proposed the Pixel-Matched Spread Function, and a simplified deconvolution arithmetic to suppress the intra-page crosstalk noise.

另外，在分析数字数据全息存储中页内串扰噪声产生的物理过程的基础上，提出了像元匹配扩散函数的概念，并提出了一个反卷积算法来抑制页内串扰噪声，计算机模拟和实验结果都证明我们提出的PMSF算法可以有效的抑制页内串扰噪声，提高恢复数据页的保真度。

Experiments and tests have been carried out based on the improved algorithm and Apriori algorithm.

论文对提出的两种改进算法与Apriori算法在实验数据集上进行测试。

Based on the classical Apriori algorithm, a novel graph mining algorithm, Apriori-Graph, is proposed.

在经典的Apriori算法的基础上，提出了一种图挖掘的新算法Apriori-Graph。

In this thesis, we research the Apriori algorithm for discovering frequent itemsets and fast algorithm for rule generation.

论文在阐述了数据挖掘的基本概念、挖掘步骤和挖掘特点的基础上，对关联规则数据挖掘理论进行了深入的研究，利用Apriori算法和规则生成快速算法分别实现了频繁项目集的发现和关联规则的生成。

The split between the two groups can hardly be papered over.

这两个团体间的分歧难以掩饰。

This approach not only encourages a greater number of responses, but minimizes the likelihood of stale groupthink.

这种做法不仅鼓励了更多的反应，而且减少跟风的可能性。