作者: Shuwen Zhang , Xuxin Gu
DOI: 10.1016/J.IJLEO.2013.03.133
关键词:
摘要: Abstract In this paper, we propose a palmprint recognition method based on the representation in feature space. The proposed seeks to represent test sample as linear combination of all training samples space and then exploits obtained perform recognition. We can implement mapping from original by using kernel functions such radial basis function (RBF). method, selection parameter is important. an automatic algorithm for selecting parameter. basic idea optimize that same class are well clustered while different classes pushed far away. criterion measures goodness space, optimal minimizing criterion. Experimental results multispectral database show more effective than 2DPCA, 2DLDA, AANNC, CRC_RLS, nearest neighbor (NN) competitive coding terms correct rate.