2DLPP:A novel method for small sample size face recognition

Ruixia Yang · Journal of Optoelectronics·laser · 2008

LPP is a local feature extraction method,which can detect the intrinsic manifold structure of data,In face recognition,it often encounters the small sample size problem.To solve that,we propose a two-dimensional locality preserving projections(2DLPP) based directly on 2D image matrices for face recognition.Experiment on Yale and Extended YaleB face database show that 2DLPP achieves better face recognition performance than LPP.

Read the paper · More papers on PaperTik