Hash Dictionary Searching Algorithm

ＨａｓｈＤｉｃｔｉｏｎａｒｙＳｅａｒｃｈｉｎｇＡｌｇｏｒｉｔｈｍＴａｎｇＪｉａｙｉ（ＤｅｐａｒｔｍｅｎｔｏｆＣｏｍｐｕｔｅｒ，Ｘｉ＇ａｎＩｎｓｔｉｔｕｔｅｏｆＰｏｓｔｓａｎｄＴｅｌｅｃｏｍｍｕｎｉｃａｔｉｏｎｓ，Ｘｒ＇ａｎ７１００６１．Ｐ．Ｒ．Ｃｈｉ...     

基于改进K-SVD字典学习的超分辨率图像重构

 针对已有算法中字典训练的时间消耗巨大的问题,提出了一种改进的基于字典学习的超分辨率图像重构算法.本文将K-SVD字典算法和高低分辨率联合生成的思想结合起来,形成新的字典训练方法,并将由该算法生成的高低分辨率字典应用于基于稀疏表示的超分辨率重构.重构仿真实验证明算法不仅有效降低了字典训练所消耗的时间,而且能够改善重构高分辨图像的质量.     

基于迭代自适应稀疏分解的雷达信号去噪

稀疏分解能有效分离信号和噪声,因此适用于信号去噪.文中构造了雷达回波稀疏表示的冗余字典,字典原子与目标回波波形匹配,基于该字典的雷达回波信号稀疏度就是目标数.针对稀疏度自适应匹配追踪算法进行低信噪比信号稀疏分解时的不足,提出了一种迭代自适应匹配追踪算法,采用规范化的残差之差作为迭代终止条件,使得稀疏分解过程能依据噪声水平自适应终止,以逐次逼近方式估计信号稀疏度,改善了稀疏分解的精度.仿真实验结果表明,该算法在低信噪比以及稀疏度未知的条件下,实现了雷达回波信号的准确稀疏分解,极大地提高了信噪比.     

Semantic dictionary based method for short text classification

The traditional short-text classification's accuracy usually highly relies on statistical feature selection. Owing to the fact that short-text has inherent defects such as short length, weak signal and less features. It is hard to avoid noise words when doing feature extension which will highly influence the accuracy of classification. In order to solve the above problem, this paper proposes a semantic dictionary method for short-text classification. The method builds a set of domain dictionary by analyzing the specific characteristics in certain field. As each word's weight in the dictionary is designed according to the correlation between the word and the category, classification accuracy has improved to some extent. Then, in order to enhance dictionary vocabulary coverage, association rules are utilized to automatically extend semantic dictionary. Finally, an experiment based on micro-blog data is conducted which shows that the method has a good effect.     

基于CANopen协议的延伸报警器研制

为了提高船舶机舱的延伸报警安全性和互联通信可靠性及实时控制性,文中提出基于LPC1768为核心采用CANopen通信协议的机舱延伸报警系统,可以进一步提高系统的可靠性、通信效率及灵活性.系统采用模块思想设计,实现了主站人工、自动切换,系统运行状态自我监测,运行状态显示和声光报警等功能,阐述了系统硬件设计、CANopen协议软件的实现过程(包括节点的对象字典和通信模型建立).结果表明该系统能够准确侦听报警信息,且结构简单.能够满足船舶延伸报警的要求、提高了延伸报警的可靠性和兼容性.     

基于匹配追踪算法的激波信号特征提取

针对时域重叠、噪声混合等情况下从多弹头激波信号中提取时延特征参数问题,利用匹配追踪(Matching pursuits,MP)算法的自适应分解特性对激波混合信号进行特征提取,并针对运行时间长和内存占用大等问题提出"粗提取—精提取"两步改进算法,通过"粗提取"获得各个激波信号的长度,触发位置等先验知识,利用"分库"、"总库"思想重构过完备原子库进行精提取。仿真结果表明,该改进算法不仅能以较高精度提取时延特征参数,而且能有效提高提取速度和减小内存占用。     

一种容差模拟电路的分级故障诊断方法

In this paper, a tolerance analog circuit fault diagnosis method based on hierarchical fault dictionary is proposed. During the simulation before test, firstly, the Worse-Case Analysis is used to get the normal characteristics output interval of the circuit under test and the output interval is saved as the first class fault dictionary, which will be used to fault detection; secondly, node-voltage sensitivity sequence is used as fault characteristics to build the second class fault dictionary for locating fault component; thirdly, based on simulation before test according to dividing the component parameters into seven segments, the third class fault dictionary is built to identify the parameter interval of components. In the fault diagnosis stage, based on the established three-class fault dictionary, fault detection, fault locating and component parameter interval identification can be realized respectively according to practical application. The proposed method can improve the efficiency of diagnosis after test and the solution will be a meaningful reference for practical applications. Finally, the simulation experiment demonstrates the effectiveness of the proposed method.     

基于K-SVD字典学习的合成图像盲检测

一幅真实的图像中噪声特性是一致的,而由多幅图像内容拼接而成的合成图像噪声特性是不一致的.本文利用这一特点,提出了一种基于K均值奇异值分解(K-SVD)字典学习的合成图像盲检测方法.该方法首先通过K-SVD算法对合成图像进行训练得到其稀疏表示字典,然后利用学习得到的字典对背景噪声进行去除,最后根据去噪前后图像对应子块的相关系数异同实现篡改区域的检测与定位.实验结果表明,该方法对于鉴别含有不同背景噪声的合成图像具有显著效果,同时,算法对JPEG压缩、重采样和模糊等后处理操作都具有较好的鲁棒性.     

基于双字典和稀疏表示的医学图像超分辨率重建

在医学影像图像处理过程中,由于成像技术和成像时间的限制,还无法获取满足诊断需求的清晰图像,这使得在现有技术和极短时间内所获取的医学病理图像需要进行超分辨率的重建处理;基于学习的图像超分辨率思想是从已建立的先验模型中重建出高频细节;在文章中,将要估计的高频信息认为是由主要高频和冗余高频两部分组成,提出了一种基于双字典学习和稀疏表示的医学图像超分辨率重建算法,由主要字典学习和冗余字典学习组成,分别渐近地恢复出主要高频细节和冗余高频细节;实验结果的数据分析和视觉效果显示,所提出双层递进方法能够恢复更多的图像细节且在性能指标上比现有的其他几种方法均有所提高。