Person re-identification by Local Maximal Occurrence representation and metric learning

TLDR

Person re-identification is crucial for automatically locating individuals in surveillance footage, yet it is hindered by the need for robust feature representations and discriminative metric learning. The authors introduce a Local Maximal Occurrence (LOMO) feature and a Cross‑view Quadratic Discriminant Analysis (XQDA) framework to address these challenges. LOMO captures horizontal occurrences of local features, incorporates Retinex transformation and scale‑invariant texture operators to mitigate illumination changes, while XQDA learns a low‑dimensional discriminant subspace and a corresponding metric, providing a practical computation method and regularization. On VIPeR, QMUL GRID, CUHK Campus, and CUHK03 datasets, the proposed approach raises rank‑1 identification rates by 2.2%, 4.88%, 28.91%, and 31.55%, respectively.

Abstract

Person re-identification is an important technique towards automatic search of a person's presence in a surveillance video. Two fundamental problems are critical for person re-identification, feature representation and metric learning. An effective feature representation should be robust to illumination and viewpoint changes, and a discriminant metric should be learned to match various person images. In this paper, we propose an effective feature representation called Local Maximal Occurrence (LOMO), and a subspace and metric learning method called Cross-view Quadratic Discriminant Analysis (XQDA). The LOMO feature analyzes the horizontal occurrence of local features, and maximizes the occurrence to make a stable representation against viewpoint changes. Besides, to handle illumination variations, we apply the Retinex transform and a scale invariant texture operator. To learn a discriminant metric, we propose to learn a discriminant low dimensional subspace by cross-view quadratic discriminant analysis, and simultaneously, a QDA metric is learned on the derived subspace. We also present a practical computation method for XQDA, as well as its regularization. Experiments on four challenging person re-identification databases, VIPeR, QMUL GRID, CUHK Campus, and CUHK03, show that the proposed method improves the state-of-the-art rank-1 identification rates by 2.2%, 4.88%, 28.91%, and 31.55% on the four databases, respectively.