摘要
为了充分利用样本的类别信息,提取出更加有效的分类特征,提出一种基于类内子空间学习的局部线性嵌入算法。该算法首先获取类内离散矩阵的子空间,然后采用类内子空间构成类间离散矩阵;为了进一步增强算法的性能,使用了结合最大边界准则与局部线性嵌入算法的目标函数。最后在人脸数据库上的实验结果表明,同其他算法相比,ISL/LLE算法具有更好的识别性能。
In order to make full use of the class information of the sample and extract more effective classification features,a local linear embedding algorithm based on intra-class subspace learning is proposed. The algorithm first obtains the subspace of the intra-class discrete matrix, and then uses the intra-class subspace to form the inter-class discrete matrix. To further enhance the performance of the algorithm, the objective function combining the maximum margin criterion and the locally linear embedding is utilized. Finally, the experimental results on the face database show that the ISL/LLE algorithm has better recognition performance than other algorithms.
引文
[1]王罡,骈俊生.一种基于人脸检测的人眼特征定位方法[J].阜阳师范学院学报(自然科学版),2010,27(1):37-39.
[2]邢慧芬,吴其林,曹骞.基于人类视觉模型和Contourlet变换的图像感知哈希算法[J].阜阳师范学院学报(自然科学版),2016(4):62-66.
[3]WOLD S,ESBENSEN K,GELAD P.Principal component analysis[J].Chemometrics&Intelligent Laboratory Systems,1987,2(1):37-52.
[4]TURK M,PENTLAND A.Eigenface for recognition[J].Journal of Cognitive Neuros-cience,1991,3(1):71-86.
[5]TREDER M S,PORBADBUGK A K,AVARVAND FS,et al.The LDA beamformer:Optimal estimation of ERP source time series using linear discriminant analysis[J].Neuroimage,2016,129:279-291.
[6]CHEN L C.An effective LDA-based time topic model to improve blog search performance☆[J].Information Processing&Management,2017,53(6):1299-1319.
[7]SWETS D,WENG J.Using discriminant eigenfeatures for image retrieval[J].IEEE Trans on Pattern Analysis&Machine Intellgence,1996,18(8):831-836.
[8]BELHUMEUR P,HESPANHA J,KRIEGMAN D.Eigenfaces vs.Fisherfaces:recognition using class specific linear projection[J].IEEE Transaction on Pattern Analysis&Machine Intelligence,1997,19(7):711-720.
[9]SUBASI A,ISMAIL G M.EEG signal classification using PCA,ICA,LDA and support vector machines[J].Expert Systems with Applications,2010,37(12):8659-8666.
[10]YE J.Characterization of a family of algorithms for generalized discriminant analysis on under sampled problems[J].Journal of Machine Learning Research,2005(6):483-502.
[11]WU X J,KITTLER J,YANG J Y,et al.A new direct LDA(D-LDA)Algorithm for Feature Extraction in Face Recognition[C]//International Conference on Pattern Recognition.2004.
[12]LI H,JIANG T,ZHANG K.Efficient and robust feature extraction by maximum margin criterion[J].IEEETransactions on Neural Networks,2006,17(1):157-165.
[13]卢桂馥,邹健,陈富春.一种求解MMC的快速算法[J].安徽工程大学学报,2014(4):57-61.
[14]江艳霞,吴腾飞,刘子渊.基于加权邻域极大边界判别式嵌入的人脸识别算法[J].计算机工程,2016,42(6):167-170.
[15]HE X F,NIYOGI P.Locality Preserving Projections,Advances in Neural lnformation Processing Systems//Thrun S,Saul L,Scholkopf B(eds.).Cambridge,USA:MIT Press,2004,16:153-160.
[16]何力,张军平,周志华.基于放大因子和延伸方向研究流形学习算法[J].计算机学报,2005,28(12):2000-2009.
[17]ROWEIS T,SAUL K.Nonlinear dimensionality reduction by locally linear embedding.[J].Science,2000,290(5500):2323-2326.
[18]孙锐,王旭,张东东,等.采用局部线性嵌入的稀疏目标跟踪方法[J].电子测量与仪器学报,2017,31(8):1312-1320.
[19]刘方原,夏克文,牛文佳.改进重构权值的局部线性嵌入算法[J].中国图象图形学报,2018,23(1):52-60.
[20]钱文光,李会民.一种相似子空间嵌入算法[J].山东大学学报(工学版),2018,48(1):8-14.
