Information geometry theory of high-dimension space and application for speaker independent continuous digit speech recognition

In the light of descriptive geometry and notions in set theory, this paper re-defines the basic elements in space such as curve and surface and so on, presents some fundamental notions with respect to the point cover based on the High-dimension space (HDS) point covering theory, finally takes points from mapping part of speech signals to HDS, so as to analyze distribution information of these speech points in HDS, and various geometric covering objects for speech points and their relationship. Besides, this paper also proposes a new algorithm for speaker independent continuous digit speech recognition based on the HDS point dynamic searching theory without end-points detection and segmentation. First from the different digit syllables in real continuous digit speech, we establish the covering area in feature space for continuous speech. During recognition, we make use of the point covering dynamic searching theory in HDS to do recognition, and then get the satisfying recognized results. At last, compared to HMM (Hidden Markov models)-based method, from the development trend of the comparing results, as sample amount increasing, the difference of recognition rate between two methods will decrease slowly, while sample amount approaching to be very large, two recognition rates all close to 100% little by little. As seen from the results, the recognition rate of HDS point covering method is higher than that of in HMM (Hidden Markov models) based method, because, the point covering describes the morphological distribution for speech in HDS, whereas HMM-based method is only a probability distribution, whose accuracy is certainly inferior to point covering.

Research of Face Location System Based on Human Vision Simulations

In this paper we present a robust face location system based on human vision simulations to automatically locate faces in color static images. Our method is divided into four stages. In the first stage we use a gauss low-pass filter to remove the fine information of images, which is useless in the initial stage of human vision. During the second and the third stages, our technique approximately detects the image regions, which may contain faces. During the fourth stage, the existence of faces in the selected regions is verified. Having combined the advantages of Bottom-Up Feature Based Methods and Appearance-Based Methods, our algorithm performs well in various images, including those with highly complex backgrounds.

一种基于亮度分析的图像分割方法

图像分割是数字图像处理领域的重要研究内容。随着数字图像处理技术的发展和相关学科的进步，图像分割在图像编辑、计算机视觉、医疗影像、遥感图像等方面都取得了良好的应用，而且具有巨大的应用潜力。 图像分割是一个既很基础又很复杂的问题。经过多年的发展，研究者提出了很多优秀的算法，取得了令人鼓舞的成就。但由于图像分割本身的不确定性和复杂性，还没有一种算法能够解决好所有的分割问题。为了使图像分割的结果更加符合用户需求，交互式的分割方法逐渐成为了图像分割算法的主流。交互方式的简化也就成为了研究者努力的目标。 另一方面，图切分技术(Graph-Cut)是应用于图像分割算法的一种重要工具。Lazy Snapping方法利用人工交互操作来提取图像信息，然后使用图切分来分割图像，取得了较好的分割效果。而且该方法的交互界面简单易用，对模糊边界有较快的响应速度，引起了研究者的重视，是一种结合人工交互与图切分技术的重要分割方法。 鉴于Lazy Snapping方法在处理色彩丰富、对比强烈的图像时出现的交互操作过多、分割容易出错的问题，本文提出了一种基于亮度分析的图像分割方法。该方法在HSV空间下，以亮度为依据来提取前景和背景的特征，取代或简化了Lazy Snapping中的人工操作步骤，然后用图切分技术来进行分割。这种方法能够提取到较准确的图像特征，有效地减轻了用户操作量，明显地改善了分割结果。

一种基于模糊集合论的手势分割算法

提出了一种基于单目的复杂环境下强抗干扰性的手势分割算法,使用模糊集合的概念来描述视频流时域和空域上的不同信息,以模糊运算作为信息加工处理的工具。定义了三个模糊集合非背景集、肤色集和模糊手势集,讨论了对模糊集合的腐蚀和膨胀运算。通过对非背景集和肤色集进行模糊运算,得到原始的模糊手势集,然后对原始的模糊手势集进行求精处理。试验结果证明,该文算法实现了对人手的精确分割,且能满足实时性要求。

基于PLSA模型的文本分割

文本分割在信息提取、文摘自动生成、语言建模、首语消解等诸多领域都有极为重要的应用·基于PLSA模型的文本分割试图使隐藏于片段内的不同主题与文本表面的词、句对建立联系·实验以汉语的整句作为基本块,尝试了多种相似性度量手段及边界估计策略,同时考虑相邻句重复的未登录词对相似值的影响,其最佳结果表明,片段边界的识别错误率为6·06%,远远低于其他同类算法·

基于LDA模型的文本分割

文本分割在信息提取、文摘自动生成、语言建模、首语消解等诸多领域都有极为重要的应用.基于LDA模型的文本分割以LDA为语料库及文本建模,利用MCMC中的Gibbs抽样进行推理,间接计算模型参数,获取词汇的概率分布,使隐藏于片段内的不同主题与文本表面的字词建立联系.实验以汉语的整句作为基本块,尝试多种相似性度量手段及边界估计策略,其最佳结果表明二者的恰当结合可以使片段边界的识别错误率远远低于其它同类算法.

MST在手写汉字切分中的应用

手写汉字切分是根据输入笔迹的空间位置关系进行汉字部件的合并切分,形成完整的汉字笔划以便进行识别处理.综合利用了汉字部件的结构位置关系和笔划的空间位置关系,根据笔划的最小生成树(minimalspanningtree,简称MST)对联机连续手写输入汉字进行切分,取得了较好的切分结果.切分的准确率超过91.6%.

基于第二代Curvelet变换的指纹预处理算法

An algorithm for the enhancement of fingerprints was presented;an algorithm for realizing the foreground/background segmentation was also discussed,based on a new mathematical transform,namely, the second generation curvelet transform.The result shows that the algorithm is effective and the transform is attractive.分析了现有的一些增强、分割算法的不足,提出了基于第二代Curvelet变换的指纹图像预处理算法.实验结果表明提出的算法简单有效,能够实现指纹图像增强的要求.

Research on extracting method of micro-scale remote sensing information combination and application in coastal zone

973 Project of China [2006CB701305]; "863" Project of China [2009AA12Z148]; National Natural Science Foundation of China [40971224]

A close phylogenetic relationship between Sipuncula and Annelida evidenced from the complete mitochondrial genome sequence of Phascolosoma esculenta

Background: There are many advantages to the application of complete mitochondrial (mt) genomes in the accurate reconstruction of phylogenetic relationships in Metazoa. Although over one thousand metazoan genomes have been sequenced, the taxonomic sampling is highly biased, left with many phyla without a single representative of complete mitochondrial genome. Sipuncula (peanut worms or star worms) is a small taxon of worm-like marine organisms with an uncertain phylogenetic position. In this report, we present the mitochondrial genome sequence of Phascolosoma esculenta, the first complete mitochondrial genome of the phylum. Results: The mitochondrial genome of P. esculenta is 15,494 bp in length. The coding strand consists of 32.1% A, 21.5% C, 13.0% G, and 33.4% T bases (AT = 65.5%; AT skew = -0.019; GC skew = -0.248). It contains thirteen protein-coding genes (PCGs) with 3,709 codons in total, twenty-two transfer RNA genes, two ribosomal RNA genes and a non-coding AT-rich region (AT = 74.2%). All of the 37 identified genes are transcribed from the same DNA strand. Compared with the typical set of metazoan mt genomes, sipunculid lacks trnR but has an additional trnM. Maximum Likelihood and Bayesian analyses of the protein sequences show that Myzostomida, Sipuncula and Annelida (including echiurans and pogonophorans) form a monophyletic group, which supports a closer relationship between Sipuncula and Annelida than with Mollusca, Brachiopoda, and some other lophotrochozoan groups. Conclusion: This is the first report of a complete mitochondrial genome as a representative within the phylum Sipuncula. It shares many more similar features with the four known annelid and one echiuran mtDNAs. Firstly, sipunculans and annelids share quite similar gene order in the mitochondrial genome, with all 37 genes located on the same strand; secondly, phylogenetic analyses based on the concatenated protein sequences also strongly support the sipunculan + annelid clade (including echiurans and pogonophorans). Hence annelid "key-characters" including segmentation may be more labile than previously assumed.

基于马尔可夫随机场模型的图像分割技术研究

本文针对基于马尔可夫随机场模型（MRF）的图像分割技术进行研究，通过深入分析马尔可夫随机场模型用于图像分割时的优缺点，提出了改进方案，将其用于单帧图像的无监督分割和动态场景下的运动目标分割。主要研究内容包括以下几部分。 第一部分详细介绍了马尔可夫随机场模型，包括邻域系统和基团的概念、初始标记场的获取、能量函数的确立和MAP估算方法。 第二部分针对噪声图像的预处理，提出一种多尺度双边滤波算法来综合不同尺度下双边滤波的去噪效果。为降低双边滤波的计算复杂性，提出一种双边滤波快速计算方法。该算法能够在去除噪声的同时较好地保留边缘。 第三部分针对MRF模型用于图像分割中遇到的过平滑问题，定义了一种间断自适应高斯马尔可夫随机场模型（DA-GMRF），提出一种基于该模型的无监督图像分割方法。利用灰度直方图势函数自动确定分类数及分割阈值，进行多阈值分割得到标记场的初始化，用Metroplis采样器算法进行标记场的优化，得到最终的分割结果。该方法考虑了平滑约束在图像边缘处的自适应性，避免了边缘处的过平滑，将其应用于无监督图像分割取得了较好的效果。 第四部分针对动态场景下的运动目标分割，提出一种基于间断自适应时空马尔可夫随机场模型的运动目标分割方法。解决了传统时空马尔可夫随机场模型不能对运动造成的显露遮挡现象进行处理问题，也克服了全局一致平滑假设造成的过平滑问题。帧差图像二值化得到初始标记场，初始标记场进行‘与’操作获得共同标记场，用Metroplis采样器算法实现共同标记场的优化。该方法既使用了平滑约束，而又保留了间断，从而使分割得到的运动目标边缘更加准确。

曲线演化图像分割和森林烟火预警系统

随着计算机技术，图像采集技术和数据存储技术等的进步，图像处理的应用领域越来越广泛。很多的应用系统是综合利用了电子，通讯和图像处理等技术而开发出来的，图像处理往往是系统的核心部分。图像分割是图像处理的核心技术，也是图像处理技术中的难点。所以研究图像分割技术具有非常重要的意义。 传统的图像分割方法有：使用模板对图像进行边缘检测等；利用滤波处理，频谱分析等数字信号处理处理技术进行分割。80年代末以来，偏微分方程方法越来越多地应用到图像分割领域中，已成为图像分割的有力工具。本文对基于偏微分方程的图像分割方法进行研究，介绍单开曲线演化分割算法，并基于Mumford-Shah模型提出一种带状目标分割方法。这种方法能将图像中的带状区域从图像中分割出来－这里假定带状区域的边界可用单值函数表示。与其它方法，如边缘检测分割，C-V模型分割和单开曲线分割相比，本文提出的方法得到的分割结果有与目标的边界吻合的更好，抗噪能力强等优点。 本文介绍了通过对可见光摄像机所拍摄图像进行分析来检测火的森林烟火预警系统。该系统是通过检测烟的存在来判断是否有火情。图像处理软件是森林烟火预警系统的核心组成部分。评价火灾预警系统性能有两个标准。一个是一旦发生火灾，预警系统能否快速地发出火警信号；另一个是在没有火情时，预警系统是否不报警，即误警率是否低。图像分割在设计图像处理算法时，主要在两个地方得到应用。在图像预处理阶段，利用单开曲线演化分割算法或带状区域的分割算法将森林区域分割出来。这样是为了在对图像进行处理时消除非森林区域中的目标对识别结果的影响，降低误警率。在图像处理阶段，利用图像分割算法将烟从图像中分割出来，准确及时报警。