2008_基于Laplace谱的图像分类_张江.pdfVIP

18 5 Vol. 18 No. 5 2008 5 COM PU ER ECHNOLOGY AND DEVELOPMEN May 2008 Laplace 张 江, 王 年, 梁 栋, 唐 俊 ( 安徽大学计算智能与信号处理教育部重点实验室, 安徽 合肥 230039) : , Laplace Laplace, SVD, , Laplace , BP SVM , , Laplace, , Laplace, , Laplace, , Laplace, : Laplace; ; ; : N911.73 :A : 1673- 629X(2008) 05- 0073- 03 Image Classification Based on Laplacian Spectra ZHANG Jiang, WANG Nian, LIANG Dong, ANG Jun (Ministry of Educat ion Key Lab. of Intelligent Computing Signal Processing,Anhui University, Hefei 230039, China) Abstract : A new algorithm for image classification using Laplacian spectrum is presented, utilizing spectrum to reflect the structural char acteristics of the images. Firstly, the eigenvalue of Laplacian matrix constructed on the feature points of images is obtained by using the method of singular value decomposition. Secondly, the high- dimensional Laplacian eigenvalue is projected onto the low - dimensional eigenspace by means of covariance matrix. Finally, image classification is perform ed by adopting back propagation algorithm and support vector machine algorithm. Experim ents of synthetic data and real image demonstrate that different classes of the sequence images have dif ferent structure, and their Laplacian spectrum are also different, whereas sim ilar classes of the sequence images have similar structure, and their Laplacian spectrum are similar, too. herefore, there have directly relationship betw een Laplacian spectrum and the graph structure, and Laplacian spectrum enables to regard as feature samples of image classification. Experimental results indicate that it is of comparatively high recognition rate to use the image characteristics denoted