Threshold-Consistent Margin Loss for Open-World Deep Metric Learning

arXiv:2307.04047v2 Announce Type: replace
Abstract: Existing losses used in deep metric learning (DML) for image retrieval often lead to highly non-uniform intra-class and inter-class representation structures across test classes and data distributions. When combined with the common practice of using a fixed threshold to declare a match, this gives rise to significant performance variations in terms of false accept rate (FAR) and false reject rate (FRR) across test classes and data distributions. We define this issue in DML as threshold …

abstract arxiv class consistent cs.cv data image loss losses match open-world practice representation retrieval test threshold type uniform world