Kernel Principal Component Analysis

Kernel Principal Component Analysis (Kernel PCA) [5,6] is a method of non-linear feature extraction, closely related to methods applied in Support Vector Machines (SVMs)[3].
A cite of code of KPCA can be seen as follows:
http://cmp.felk.cvut.cz/cmp/software/stprtool/manual/kernels/extraction/list/kpca.html