Distance metric learning (DML) has been successfully applied to object
classification, both in the standard regime of rich training data and in the
few-shot scenario, where each category is represented by only few examples. In
this work, we propose a new method for DML, featuring a joint learning of the
embedding space and of the data distribution of the training categories, in a
single training process. Our method improves upon leading algorithms for
DML-based object classification. Furthermore, it opens the door for a new task
in Computer Vision - a few-shot object detection, since the proposed DML
architecture can be naturally embedded as the classification head of any
standard object detector. In numerous experiments, we achieve state-of-the-art
classification results on a variety of fine-grained datasets, and offer the
community a benchmark on the few-shot detection task, performed on the
Imagenet-LOC dataset. The code will be made available upon acceptance.
%0 Generic
%1 citeulike:14614669
%A xxx,
%D 2018
%K classification detection finegrained metric zero\_shot
%T RepMet: Representative-based metric learning for classification and one-shot object detection
%U http://arxiv.org/abs/1806.04728
%X Distance metric learning (DML) has been successfully applied to object
classification, both in the standard regime of rich training data and in the
few-shot scenario, where each category is represented by only few examples. In
this work, we propose a new method for DML, featuring a joint learning of the
embedding space and of the data distribution of the training categories, in a
single training process. Our method improves upon leading algorithms for
DML-based object classification. Furthermore, it opens the door for a new task
in Computer Vision - a few-shot object detection, since the proposed DML
architecture can be naturally embedded as the classification head of any
standard object detector. In numerous experiments, we achieve state-of-the-art
classification results on a variety of fine-grained datasets, and offer the
community a benchmark on the few-shot detection task, performed on the
Imagenet-LOC dataset. The code will be made available upon acceptance.
@misc{citeulike:14614669,
abstract = {{Distance metric learning (DML) has been successfully applied to object
classification, both in the standard regime of rich training data and in the
few-shot scenario, where each category is represented by only few examples. In
this work, we propose a new method for DML, featuring a joint learning of the
embedding space and of the data distribution of the training categories, in a
single training process. Our method improves upon leading algorithms for
DML-based object classification. Furthermore, it opens the door for a new task
in Computer Vision - a few-shot object detection, since the proposed DML
architecture can be naturally embedded as the classification head of any
standard object detector. In numerous experiments, we achieve state-of-the-art
classification results on a variety of fine-grained datasets, and offer the
community a benchmark on the few-shot detection task, performed on the
Imagenet-LOC dataset. The code will be made available upon acceptance.}},
added-at = {2019-02-27T22:23:29.000+0100},
archiveprefix = {arXiv},
author = {xxx},
biburl = {https://www.bibsonomy.org/bibtex/220771ee1efecc25d3dc976305399744d/nmatsuk},
citeulike-article-id = {14614669},
citeulike-linkout-0 = {http://arxiv.org/abs/1806.04728},
citeulike-linkout-1 = {http://arxiv.org/pdf/1806.04728},
day = 15,
eprint = {1806.04728},
interhash = {f44af82921b7ec6685266b950b99f28c},
intrahash = {20771ee1efecc25d3dc976305399744d},
keywords = {classification detection finegrained metric zero\_shot},
month = jun,
posted-at = {2018-07-14 21:41:15},
priority = {5},
timestamp = {2019-02-27T22:23:29.000+0100},
title = {{RepMet: Representative-based metric learning for classification and one-shot object detection}},
url = {http://arxiv.org/abs/1806.04728},
year = 2018
}