Kernelized laplacian collaborative representation based classifier for face recognition

Abstract

A recently proposed sparse representation based classifier, called collaborative representation based classification with regularized least square (CRC-RLS), has attracted notable attention. The extensive experiments demonstrate that the CRC-RLS technique has less complexity than traditional sparse representation based classifier (SRC) but results in better classification performance. However, the existing SRC-like approaches fail to consider the manifold structure of the data space. It has been shown that the manifold information of the data is important for discrimination. The paper presents a more effective classification scheme, termed kernelized laplacian collaborative representation based classifier (KLCRC) for face recognition. KLCRC explicitly take into account the nonlinear distribution and local manifold structure of the data. The extensive experimental results over several standard face databases have demonstrated the effectiveness of the proposed algorithm. © Springer International Publishing 2013.