Learning Incremental Triplet Margin for Person Re-identification
This work addresses the challenge of matching people across non-overlapping cameras, offering an incremental improvement in metric learning for person re-identification.
The paper tackles the problem of person re-identification by proposing a multi-stage training strategy that learns incremental triplet margins, resulting in improved performance on standard datasets like Market-1501, CUHK03, and DukeMTMCreID, with concrete gains over most existing state-of-the-art methods.
Person re-identification (ReID) aims to match people across multiple non-overlapping video cameras deployed at different locations. To address this challenging problem, many metric learning approaches have been proposed, among which triplet loss is one of the state-of-the-arts. In this work, we explore the margin between positive and negative pairs of triplets and prove that large margin is beneficial. In particular, we propose a novel multi-stage training strategy which learns incremental triplet margin and improves triplet loss effectively. Multiple levels of feature maps are exploited to make the learned features more discriminative. Besides, we introduce global hard identity searching method to sample hard identities when generating a training batch. Extensive experiments on Market-1501, CUHK03, and DukeMTMCreID show that our approach yields a performance boost and outperforms most existing state-of-the-art methods.