Robust face recognition via sparse reconstruction vector

摘要

Recognition rate and recognition time of algorithms are the major issues in face recognition. There are several types of algorithm used for face recognition such as eigenface or fisherface. Face recognition based sparse representation is successfully used in this area recently. In this work, both recognition rate and recognition time of Principal Component Analysis-PCA (eigenface) and Sparse Representation-based Classification-SRC are compared. SRC algorithm is modified for a high recognition rate. The simulation results show that modified SRC algorithm is indisputably superior to PCA algorithm for performance of recognition rate and algorithm time if the training set has enough number of samples.