GABOR-BASED TENSOR LOCAL DISCRIMINANT EMBEDDING AND ITS APPLICATION ON PALMPRINT RECOGNITION

摘要

In this paper, a novel feature extraction method called Gabor-based tensor local discriminantnembedding (GTLDE) is proposed. GTLDE first gets the high-order statisticninformation by using a biologically inspired hierarchical model, and then tensor localndiscriminant embedding (TLDE) is carried out to extract the discriminant features ofnthe image for recognition task. The method we proposed is not only robust to localntranslation and scale variations, but also has high distinguishing ability. More importantly,nour method can achieve high accuracy with a small number of training samples.n∗Corresponding author.n327n328 L. Cui et al.nExperimental results on PolyU-II palmprint database demonstrate the effectiveness ofnthe method we proposed