Rethinking preventing class-collapsing in metric learning
with margin-based losses
Elad Levi1 , Tete Xiao2 , Xiaolong Wang 3 , Trevor Darrell1,2
1
Nexar, 2 UC Berkeley, 3 UC San Diego
Abstract sual appearance. Pushing all these modes to a single point
in the embedding space requires the network to memorize
Metric learning seeks perceptual embeddings where vi- ...
附件列表