Person Re-Identification using Kernel-based Metric Learning Methods

Abstract. Re-identification of individuals across camera networks with limited or no overlapping fields of view remains challenging in spite of significant research efforts. In this paper, we propose the use, and ex-tensively evaluate the performance, of four alternatives for re-ID clas-sification: regularized Pairwise Constrained Component Analysis, ker-nel Local Fisher Discriminant Analysis, Marginal Fisher Analysis and a ranking ensemble voting scheme, used in conjunction with different sizes of sets of histogram-based features and linear, χ2 and RBF-χ2 kernels. Comparisons against the state-of-art show significant improvements in performance measured both in terms of Cumulative Match Characteris-tic curves (CMC) and Proportion of Uncertainty Removed (PUR) scores on the challenging VIPeR, iLIDS, CAVIAR and 3DPeS datasets.