A linear discriminant analysis framework based on random subspace for face recognition

Linear discriminant analysis (LDA) often suffers from the small sample size problem when dealing with high-dimensional face data. Random subspace can effectively solve this problem by random sampling on face features. However, it remains a problem how to construct an optimal random subspace for discriminant analysis and perform the most efficient discriminant analysis on the constructed random subspace. In this paper, we propose a novel framework, random discriminant analysis (RDA), to handle this problem. Under the most suitable situation of the principal subspace, the optimal reduced dimension of the face sample is discovered to construct a random subspace where all the discriminative information in the face space is distributed in the two principal subspaces of the within-class and between-class matrices. Then we apply Fisherface and direct LDA, respectively, to the two principal subspaces for simultaneous discriminant analysis. The two sets of discriminant analysis features from dual principal subspaces are first combined at the feature level, and then all the random subspaces are further integrated at the decision level. With the discriminating information fusion at the two levels, our method can take full advantage of useful discriminant information in the face space. Extensive experiments on different face databases demonstrate its performance.     

基于邻域信息的图像分割技术

提出一种基于图像邻域信息的分割方法.首先,根据像素点邻域信息得到高维特征向量;然后采用典型相关分析(CCA)改进线性判别分析(LDA)中的变换矩阵,使得特征向量的降维具有自适应性;最后用最近邻法对降维后的特征向量进行分类,从而实现了图像的分割.试验中,选取人脸图像分割来验证该方法,结果显示出其具有良好的分类效果.     

基于自适应邻域选择的局部判别投影算法

在多模数据分类中,使用局部Fisher判别分析和边界Fisher分析方法构建邻域不能充分反映流形学习对邻域的要求.为此,提出一种基于自适应邻域选择的局部判别投影算法.采用自适应方法扩大或者缩小近邻系数k,以构建邻域,从而保持局部线性结构,揭示流形的内在几何结构,利用局部化方法使得投影空间中同类近邻样本尽量紧凑、异类近邻样本尽量分开.在ORL和YALE入脸数据库中进行实验,结果表明,在不同训练样本个数下,该算法均能获得较高的识别率.     

基于局部中频Gabor滤波器组的人脸识别

针对传统Gabor滤波器组在人脸识别过程中特征提取时间长、计算量大的问题,提出一种利用局部Gabor滤波器组进行人脸图像中频特征提取的方法.选择中频带的Gabor滤波器构造局部中频Gabor滤波器组;提取局部Gabor中频特征;采用线性判别分析法(linear discriminate analysis,LDA)进一步提取Fisher特征,得到图像的Gabor+ Fisher特征,利用最近邻法进行人脸图像识别.基于ORL和AR人脸库的实验结果表明,基于此局部Gabor滤波器组的人脸识别方法较传统的Gabor特征提取方法降低了40％的特征维数,加快了特征提取速度,提高了人脸识别率.     

融合显著信息的LDA极光图像分类

美丽的极光形态各异,不同形态的极光蕴含不同的物理意义,所以研究极光图像的分类具有重要的科学价值.在LDA(latent Dirichlet allocation)模型基础上提出了一种融合显著信息的LDA 方法(LDA with saliencyinformation,简称SI-LDA),利用极光图像的谱残差(spectral residual,简称SR)显著信息生成视觉字典,加强极光图像的语义信息,并将其用于极光图像的特征表示.最后,利用SVM分类器对极光图像进行分类.实验结果表明,所提出的算法获得了良好的分类结果.     

基于弹性网格的西夏文字识别

随着国内外对西夏学研究的不断深入,收藏于世界各地的大批西夏古籍文献通过影印方式陆续出版。如何将这些西夏古籍文献数字化、文本化则有着极其重要的意义。该文采用弹性网格方法及线性判别分析(Linear Discriminant Analysis,LDA)方法对西夏文字识别进行了研究。首先对西夏影印文献进行预处理、细化,然后根据西夏文字笔画分布构造非均匀的弹性网格,将弹性网格分别作用于西夏文字的四个方向分量上,统计像素点在网格内的概率分布作为特征,最后使用LDA方法对提取的特征降维处理。对240类共9 600个西夏文字做4重交叉验证,平均识别率可达87.99%,实验表明该方法是有效的。     

OPTIMIZATION AND NEURAL MODELLING OF PULSE COMBUSTORS FOR DRYING APPLICATIONS

ABSTRACT

Results of investigations of a valved pulse combustor to choose optimal geometry, which covered measurements of the flow rates of air and fuel, pressure oscillations, including pressure amplitude and frequency and flue gas composition are presented in the paper. Experimental studies compsiring the operation of the pulse combustor coupled with a drying chamber and working separately are described. It was found that coupling of the pulse combustor with a drying chamber had no significant effect on the pulse combustion process. Smoother runs of pressure oscillations in the combustion chamber, lower noise level and slightly higher NO_x emission were observed. The velocity flow field inside the drying chamber was measured by LDA technique. Results confirmed a complex character of pulsating flow in the chamber. A large experimental data set obtained from measurements enabled developing a neural model of pulse combustion process. Artificial neural networks were trained to predict amplitudes and frequencies of pressure oscillations, temperatures in the combustion chamber and emission of toxic substances. An excellent mapping performance of the developed neural models was obtained. Due to complex character of the pulse combustion process, the application of artificial neural networks seems to be the best way to predict inlet parameters of a drying agent produced by the pulse combustor     

LARGE-EDDY SIMULATION OF FLOW AROUND CYLINDER ARRAYS AT A SUBCRITICAL REYNOLDS NUMBER

The complex three-dimensional turbulent flows around a cylinder array with four cylinders in an in-line square configuration at a subcritical Reynolds number of 1.5 × 10^4 with the spacing ratio at L/D = 1.5 and 3.5 were investigated using the Large Eddy Simulation （LES）. The full field vorticity and velocity distributions as well as turbulent quantities were calculated in detail and the near wake structures were presented. The results show that the bi-stable flow nature was observed at L/D = 1.5 and distinct vortex shedding of the upstream cylinders occurred at L/D = 3.5 at Re = 1.5 × 10^4. The techniques of Laser Doppler Anemometry （LDA） and Digital Particle Image Velocimetry （DPIV） are also employed to validate the present LES method. The results show that the numerical predictions are in excellent agreement with the experimental measurements. Therefore, the full field instantaneous and mean quantities of the flow field, velocity field and vorticity field can be extracted from the LES results for further study of the complex flow characteristics.     

Color space normalization: Enhancing the discriminating power of color spaces for face recognition

This paper presents the concept of color space normalization (CSN) and two CSN techniques, i.e., the within-color-component normalization technique (CSN-I) and the across-color-component normalization technique (CSN-II), for enhancing the discriminating power of color spaces for face recognition. Different color spaces usually display different discriminating power, and our experiments on a large scale face recognition grand challenge (FRGC) problem reveal that the RGB and XYZ color spaces are weaker than the I₁I₂I₃, YUV, YIQ, and LSLM color spaces for face recognition. We therefore apply our CSN techniques to normalize the weak color spaces, such as the RGB and the XYZ color spaces, the three hybrid color spaces XGB, YRB and ZRG, and 10 randomly generated color spaces. Experiments using the most challenging FRGC version 2 Experiment 4 with 12,776 training images, 16,028 controlled target images, and 8,014 uncontrolled query images, show that the proposed CSN techniques can significantly and consistently improve the discriminating power of the weak color spaces. Specifically, the normalized RGB, XYZ, XGB, and ZRG color spaces are more effective than or as effective as the I₁I₂I₃, YUV, YIQ and LSLM color spaces for face recognition. The additional experiments using the AR database validate the generalization of the proposed CSN techniques. We finally explain why the CSN techniques can improve the recognition performance of color spaces from the color component correlation point of view.     

基于改进用户兴趣模型的个性化图书推荐算法

为了有效提高推荐算法的精确度,提出了一种适用于个性化图书推荐的改进隐含狄利克雷分配(Latent Dirichlet Allocation,LDA)用户兴趣模型。首先在借阅者-借阅者评分矩阵的基础上,通过增加借阅者特征信息相似度计算和借阅者-图书属性相似度计算,对图书内容相似度计算方法进行了改进。然后采用LDA主题挖掘模型来实现个性化图书推荐,并给出了相应的参数估计过程。实验结果显示:相比传统算法,提出的算法具有较高的准确度,能有效对图书进行挖掘,为借阅者推荐个性化和潜在感兴的书籍。