Large scale metric learning from equivalence constraints

TLDR

Large‑scale metric learning is hindered by computationally heavy optimization and the infeasibility of fully supervised labels, making equivalence constraints a more practical supervision signal. This work proposes a scalable Mahalanobis metric learning approach that learns from equivalence constraints to address these scalability and supervision challenges. The method uses a statistical inference framework that learns a distance metric directly from equivalence constraints without complex optimization iterations. It is orders of magnitude faster than existing methods and achieves state‑of‑the‑art performance on diverse benchmarks, including face recognition, unseen object matching, and person re‑identification.

Abstract

In this paper, we raise important issues on scalability and the required degree of supervision of existing Mahalanobis metric learning methods. Often rather tedious optimization procedures are applied that become computationally intractable on a large scale. Further, if one considers the constantly growing amount of data it is often infeasible to specify fully supervised labels for all data points. Instead, it is easier to specify labels in form of equivalence constraints. We introduce a simple though effective strategy to learn a distance metric from equivalence constraints, based on a statistical inference perspective. In contrast to existing methods we do not rely on complex optimization problems requiring computationally expensive iterations. Hence, our method is orders of magnitudes faster than comparable methods. Results on a variety of challenging benchmarks with rather diverse nature demonstrate the power of our method. These include faces in unconstrained environments, matching before unseen object instances and person re-identification across spatially disjoint cameras. In the latter two benchmarks we clearly outperform the state-of-the-art.