基于局部线性嵌入的视频镜头变换检测

介绍了一种新的视频镜头变换类型检测方法,利用局部线性嵌入算法提取镜头变换的一维流形特征,用K近邻分类器确定渐变镜头变换的类型.实验表明,本算法镜头变换类型的识别率为97.5%以上.     

Citrus Essential Oil Deterpenation by Liquid‐Liquid Extraction

The object of the present work is to evaluate the possibility of carrying out the deterpenation of the citrus essential oil by using a mixture of ethanol and water as a solvent for liquid‐liquid extraction. Liquid‐liquid equilibrium has been determined for limonene + linalool + ethanol + water quaternary system at 298.15 K. Partitioning and selectivity of extraction were analyzed. Experimental data were correlated using the UNIQUAC and NRTL equations and the energetic parameters of these models at this temperature are determined.     

基于最小剩余方差的LLE改进算法

针对局部线性嵌入算法LLE算法在当流形呈卷曲状、两个曲面间距离比较小时,可能造成流形结构在重构过程的扭曲,以及近邻个数K,降维维数D值选择过程中没有一致的标准导致的降维效果下降等问题,提出一种基于改进距离的并根据剩余方差来智能选取参数值的LLE算法。该算法通过引入新的距离度量公式来替代原有算法中的欧氏距离,并根据K,D值引入剩余方差来评估高维数据结构嵌入到低维空间的效果好坏。该方法在UCI数据集和yale人脸库中进行了验证。MATLAB编程实验结果表明,该方法在选取参数值和识别率方面比传统方法有更好的性能。     

基于改进流形学习的数据分类算法

针对高维数据分类问题的特点,提出一种基于改进型局部线性嵌入LLE(Locally Linear Embedding)算法的数据降维算法,结合支持向量机SVM(Support Vector Machine)算法实现数据分类。首先,通过LLE算法降维后的数据集,按照数据集内的离差最小化,数据集间的离差最大化的原则,计算得到最优化邻近点个数;其次,将最优邻近点个数所得的降维数据作为最优结果,按一定比例选取训练集,输入SVM算法建立数据分类器;最后,将测试集输入训练完成的分类器中,实现最优化数据分类。选取Iris flower,Yale等多类数据集与传统算法进行对比实验,验证算法的可行性。实验结果表明:所提出的算法可以有效地完成数据分类,针对低维数据和高维数据分类问题具有较好的适用性和优越性,在人脸检测中也取得较好的结果。     

一种基于局部流形结构的无监督特征学习方法(英文)

Unsupervised feature selection is fundamental in statistical pattern recognition, and has drawn persistent attention in the past several decades. Recently, much work has shown that feature selection can be formulated as nonlinear dimensionality reduction with discrete constraints. This line of research emphasizes utilizing the manifold learning techniques, where feature selection and learning can be studied based on the manifold assumption in data distribution. Many existing feature selection methods such as Laplacian score, SPEC(spectrum decomposition of graph Laplacian), TR(trace ratio) criterion, MSFS(multi-cluster feature selection) and EVSC(eigenvalue sensitive criterion) apply the basic properties of graph Laplacian, and select the optimal feature subsets which best preserve the manifold structure defined on the graph Laplacian. In this paper, we propose a new feature selection perspective from locally linear embedding(LLE), which is another popular manifold learning method. The main difficulty of using LLE for feature selection is that its optimization involves quadratic programming and eigenvalue decomposition, both of which are continuous procedures and different from discrete feature selection. We prove that the LLE objective can be decomposed with respect to data dimensionalities in the subset selection problem, which also facilitates constructing better coordinates from data using the principal component analysis(PCA) technique. Based on these results, we propose a novel unsupervised feature selection algorithm,called locally linear selection(LLS), to select a feature subset representing the underlying data manifold. The local relationship among samples is computed from the LLE formulation, which is then used to estimate the contribution of each individual feature to the underlying manifold structure. These contributions, represented as LLS scores, are ranked and selected as the candidate solution to feature selection. We further develop a locally linear rotation-selection(LLRS) algorithm which extends LLS to identify the optimal coordinate subset from a new space. Experimental results on real-world datasets show that our method can be more effective than Laplacian eigenmap based feature selection methods.     

Optimization of ethanol extraction and further purification of isoflavones from soybean sprout cotyledon

Isoflavones from cotyledons of soybean sprouts were extracted with aqueous ethanol and further concentrated to obtain a product with a high concentration of isoflavone. The ethanol concentration, extraction time and reaction temperature were optimized by using response surface methodology (RSM). Isoflavones in aqueous ethanol were concentrated by a three-step procedure comprised of solid phase extraction (SPE) with Diaion HP-20 and Amberlite-XAD-2 adsorption columns, acid hydrolysis, and liquid–liquid extraction. The maximum amount of isoflavone in aqueous ethanol extracts (11.6 mg/g solid) was obtained when isoflavones in cotyledons (2.18 mg/ g solid) were extracted with 80–90% (v/v) aqueous ethanol above 90 °C for more than 100 min. The isoflavone extracts, obtained by SPE with a Diaion HP-20 column contained 100 mg/g solid. The liquid–liquid extraction (LLE) with ethyl ether further concentrated the extracts up to 229 mg/g solid, retaining 63% of the initial isoflavones.     

两种前处理方法分析竹醋液挥发性成分的比较

为探明用于熏液的竹醋液挥发性成分,分别采用吹扫捕集-热脱附法(purge and trap-thermal desorption,P＆T-TD)和液液分配萃取(liquid-liquid extraction,LLE)法对竹醋液中挥发性成分进行富集,以气相色谱-质谱联用法(gas chromatography-mass spectrometry,GC-MS)进行定性检测,用面积归一化法确定各组分的相对含量。结果表明：两种方法共鉴定竹醋液挥发性成分--酸类、酚类、酮类、酯类、醛类等化合物共50种,其中P＆T-TD-GC-MS鉴定出28种、LLE-GC-MS鉴定出29种,相对含量在2%以上的主要成分为乙酸甲酯、丙酸甲酯、乙酸、1-羟基-2-丁酮、环戊酮、糠醛、苯酚、甲基环戊烯醇酮和2,6-二甲氧基苯酚。两种方法获得的竹醋液挥发性成分差异较大,具有一定的互补性,相结合应用两种方法能够较全面地检测出竹醋液的挥发性成分。     

流形学习算法在中文文本分类中的应用

传统的流形学习局部线性嵌入 (locally linear embedding, LLE) 算法通过欧氏距离来选择邻域,如果数据集选自多个类别,这种距离度量方法无法得到正确的邻域关系。本研究提出一种改进的局部线性嵌入 (modified LLE,MLLE) 算法,该算法通过改进距离矩阵,使得类间的距离大、类内的距离小,从而使得邻域的选择尽量在一个类中。将MLLE算法应用到中文文本分类中,结果表明:与传统的算法比较,MLLE在分类结果可视化效果和识别率等方面都有显著提高。     

基于Gabor小波和局部线性嵌入的人脸识别

提出了一种新的人脸识别算法。该算法采用Gabor小波和一种新颖的方式来提取人脸特征,利用局部线性嵌入（Locally Linear Embedding,LLE）算法来实现数据的非线性降维处理,最后训练基于欧式距离的最近邻分类器进行分类判决。在ORL人脸库中与PCA方法、Gabor小波+PCA方法和直接的LLE算法进行了实验比较,实验结果表明,提出的Gabor小波+LLE的方法具有更优的性能。     

烷烃-芳烃-N-甲酰吗啉体系液液平衡

用液液平衡釜测定了等温下正庚烷苯ＮＦＭ、甲基环己烷苯ＮＦＭ、正庚烷甲苯ＮＦＭ和甲基环己烷甲苯ＮＦＭ等４组三元体系的液液平衡数据,用ＮＲＴＬ和ＵＮＩＱＵＡＣ模型进行了关联,求得各三元体系的褶点数据,并对ＮＦＭ溶剂存在下烷烃和芳烃的选择性进行了比较。