RepMet: Representative-based metric learning for classification and one-shot object detection
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 a few examples. In this work, we propose a new method for DML that simultaneously learns t...
Main Authors: | , , , , , , , |
---|---|
Format: | Report |
Language: | unknown |
Published: |
arXiv
2018
|
Subjects: | |
Online Access: | https://dx.doi.org/10.48550/arxiv.1806.04728 https://arxiv.org/abs/1806.04728 |
id |
ftdatacite:10.48550/arxiv.1806.04728 |
---|---|
record_format |
openpolar |
spelling |
ftdatacite:10.48550/arxiv.1806.04728 2023-05-15T16:01:18+02:00 RepMet: Representative-based metric learning for classification and one-shot object detection Karlinsky, Leonid Shtok, Joseph Harary, Sivan Schwartz, Eli Aides, Amit Feris, Rogerio Giryes, Raja Bronstein, Alex M. 2018 https://dx.doi.org/10.48550/arxiv.1806.04728 https://arxiv.org/abs/1806.04728 unknown arXiv arXiv.org perpetual, non-exclusive license http://arxiv.org/licenses/nonexclusive-distrib/1.0/ Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences Preprint Article article CreativeWork 2018 ftdatacite https://doi.org/10.48550/arxiv.1806.04728 2022-04-01T09:35:19Z 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 a few examples. In this work, we propose a new method for DML that simultaneously learns the backbone network parameters, the embedding space, and the multi-modal distribution of each of the training categories in that space, in a single end-to-end training process. Our approach outperforms state-of-the-art methods for DML-based object classification on a variety of standard fine-grained datasets. Furthermore, we demonstrate the effectiveness of our approach on the problem of few-shot object detection, by incorporating the proposed DML architecture as a classification head into a standard object detection model. We achieve the best results on the ImageNet-LOC dataset compared to strong baselines, when only a few training examples are available. We also offer the community a new episodic benchmark based on the ImageNet dataset for the few-shot object detection task. Report DML DataCite Metadata Store (German National Library of Science and Technology) |
institution |
Open Polar |
collection |
DataCite Metadata Store (German National Library of Science and Technology) |
op_collection_id |
ftdatacite |
language |
unknown |
topic |
Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences |
spellingShingle |
Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences Karlinsky, Leonid Shtok, Joseph Harary, Sivan Schwartz, Eli Aides, Amit Feris, Rogerio Giryes, Raja Bronstein, Alex M. RepMet: Representative-based metric learning for classification and one-shot object detection |
topic_facet |
Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences |
description |
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 a few examples. In this work, we propose a new method for DML that simultaneously learns the backbone network parameters, the embedding space, and the multi-modal distribution of each of the training categories in that space, in a single end-to-end training process. Our approach outperforms state-of-the-art methods for DML-based object classification on a variety of standard fine-grained datasets. Furthermore, we demonstrate the effectiveness of our approach on the problem of few-shot object detection, by incorporating the proposed DML architecture as a classification head into a standard object detection model. We achieve the best results on the ImageNet-LOC dataset compared to strong baselines, when only a few training examples are available. We also offer the community a new episodic benchmark based on the ImageNet dataset for the few-shot object detection task. |
format |
Report |
author |
Karlinsky, Leonid Shtok, Joseph Harary, Sivan Schwartz, Eli Aides, Amit Feris, Rogerio Giryes, Raja Bronstein, Alex M. |
author_facet |
Karlinsky, Leonid Shtok, Joseph Harary, Sivan Schwartz, Eli Aides, Amit Feris, Rogerio Giryes, Raja Bronstein, Alex M. |
author_sort |
Karlinsky, Leonid |
title |
RepMet: Representative-based metric learning for classification and one-shot object detection |
title_short |
RepMet: Representative-based metric learning for classification and one-shot object detection |
title_full |
RepMet: Representative-based metric learning for classification and one-shot object detection |
title_fullStr |
RepMet: Representative-based metric learning for classification and one-shot object detection |
title_full_unstemmed |
RepMet: Representative-based metric learning for classification and one-shot object detection |
title_sort |
repmet: representative-based metric learning for classification and one-shot object detection |
publisher |
arXiv |
publishDate |
2018 |
url |
https://dx.doi.org/10.48550/arxiv.1806.04728 https://arxiv.org/abs/1806.04728 |
genre |
DML |
genre_facet |
DML |
op_rights |
arXiv.org perpetual, non-exclusive license http://arxiv.org/licenses/nonexclusive-distrib/1.0/ |
op_doi |
https://doi.org/10.48550/arxiv.1806.04728 |
_version_ |
1766397221905891328 |