Person Re-Identification by Support Vector Ranking

TLDR

Person re‑identification matches observations of individuals across disjoint camera views, a task that becomes difficult in busy scenes due to many potential matches and appearance changes from lighting, angles, and poses, and existing approaches rely on discriminative feature extraction followed by distance‑based template matching. This work reformulates person re‑identification as a ranking problem, learning a subspace that places the true match at the top and introduces an Ensemble RankSVM to address scalability limitations of prior SVM‑based ranking methods. By shifting from absolute scoring to relative ranking, the authors implement Ensemble RankSVM and conduct extensive experiments to compare the ranking approach against template matching and classification models. The proposed ranking model markedly reduces memory usage, achieving comparable or better performance while being far more scalable than existing methods.

Abstract

Solving the person re-identification problem involves matching observation s of individuals across disjoint camera views. The problem becomes particularly hard in a busy public scene as the number of possible matches is very high. This is further compounded by significant appearance changes due to varying lighting conditions, vie wing angles and body poses across camera views. To address this problem, existing approaches focus on extracting or learning discriminative features followed by template matching using a distance measure. The novelty of this work is that we reformulate the person reidentification problem as a ranking problem and learn a subspace where th e potential true match is given highest ranking rather than any direct distance measure. By doing so, we convert the person re-identification problem from an absolute scoring p roblem to a relative ranking problem. We further develop an novel Ensemble RankSVMto overcome the scalability limitation problem suffered by existing SVM-based ranking methods. This new model reduces significantly memory usage therefore is much more scalable, whilst maintaining high-level performance. We present extensive experiments to demonstrate the performance gain of the proposed ranking approach over existing template matching and classification models.