Unconstrained face verification by subspace similarity metric learning

Unconstrained face verification, whose task is to predict whether pairs of images are from the same person or not, is always a challenging problem because of complex background, illumination, face pose, and facial expression, etc. In this paper, we propose an efficient similarity metric learning method, based on subspace learning, for face verification. Compared with relevant state-of-the-art work, this method improves the efficiency of learning the cosine similarity while keeping effectiveness. For the optimization problem, we solve it with the advanced L-BFGS optimization algorithm. We perform extensive experiments on the LFW data set using three descriptors: LBP, OCLBP, SIFT. Experiments show that our proposed method achieves the very great results.