Few-Shot Detection

Contrastive Loss

A loss function used in metric learning that minimizes the distance between pairs of examples from the same class (anchors and positives) while maximizing the distance to examples from different classes (negatives).

← Terug