Deep Cosine Metric Learning for Person Re-Identification
This addresses the problem of person re-identification for computer vision applications, offering a simpler alternative to complex sampling strategies in metric learning, though it is incremental.
The paper tackles person re-identification by proposing a method to learn a feature space optimized for cosine similarity through a re-parametrization of softmax classification, achieving competitive results on large-scale datasets and better generalization than triplet loss.
Metric learning aims to construct an embedding where two extracted features corresponding to the same identity are likely to be closer than features from different identities. This paper presents a method for learning such a feature space where the cosine similarity is effectively optimized through a simple re-parametrization of the conventional softmax classification regime. At test time, the final classification layer can be stripped from the network to facilitate nearest neighbor queries on unseen individuals using the cosine similarity metric. This approach presents a simple alternative to direct metric learning objectives such as siamese networks that have required sophisticated pair or triplet sampling strategies in the past. The method is evaluated on two large-scale pedestrian re-identification datasets where competitive results are achieved overall. In particular, we achieve better generalization on the test set compared to a network trained with triplet loss.