Penalized collaborative representation based classification for face recognition

摘要

The collaborative representation classification (CRC) exhibits superiority in both accuracy and computational efficiency. However, when representing the test sample by a linear combination of the training samples, the CRC does not account for the following: the probability of the test sample being from the same class as the training sample far from it is small. In this paper, we propose the algorithm, Penalized Collaborative Representation (PCR), which first uses the original collaborative representation to compute the distance between each training and test sample, and then treats these distances as penalized coefficients to design the penalized collaborative representation. The experimental results on multiple face databases show that our classifier, designed according PCR, has a very satisfactory classification performance.