CVPR2019論文サマリ

tag: dml

A Theoretically Sound Upper Bound on the Triplet Loss for Improving the Efficiency of Deep Distance Metric Learning
by: Takaya Yamazoe
metric learning DML triplet loss clustering