Meanings of triplet loss in English

In addition, based on the above dual hard batch construction, an ℓ22 triplet loss function is built for optimizing the training model.

Unlike triplet loss that is calculated based on triplets, SN loss is derived from K-nearest neighbors (SNs) of anchor samples.

Triplet loss and softmax loss are the two losses that are extensively used, both of which, however, have limitations.

Triplet loss optimizes the model to produce features with which samples from the same class have higher similarity than those from different classes.