Kernel collaborative representation for face recognition
Jia Zhao, Yanjiang Wang, Baodi Liu · 2014
Recently, sparse representation based classification (SRC) and collaborative representation based classification (CRC) have achieved superior performance in pattern classification. Collaborative representation based classification with regularized least square (CRC_RLS) which uses l2-norm is a very simple yet much more efficient scheme for face recognition (FR). Motivated by the fact that kernel representation is a powerful tool in discovering nonlinear structure of complex data, which may reduce the feature quantization error and boost the recognition performance, we propose Kernel Collaborative Representation based Classification (KCRC) which extends the CRC_RLS scheme to the kernel space. Compared with SRC and CRC_RLS, KCRC can greatly reduce the feature reconstruction error and learn more discriminative sparse codes for face recognition. Extensive experimental results show that the performance of KCRC outperforms the performance of support vector machine and CRC_RLS, and achieves superior performance for face recognition on several benchmark datasets.