Re-ranking with ranking-reflected similarity for person re-identification

Abstract

A recent concern for person re-identification (Re-ID) is re-ranking after initial results to improve Re-ID accuracy. In this paper, we propose a novel re-ranking method using a ranking-reflected metric to measure the similarity between the ordered set of K-nearest neighbors (OKNN) of a probe and that of a gallery. The proposed metric for ranking-reflected similarity (RSS) reflects the ranking of the shared elements between the two OKNNs. Using RSS, a re-ranking procedure is proposed that prioritizes galleries having neighbors similar to a probe's neighbor in the perspective of ranking order. In the experiment, we show that the proposed method improves the Re-ID accuracy by add-on to the state-of-the-art methods. (C) 2019 Published by Elsevier B.V.