Learning with Memory-based Virtual Classes for Deep Metric Learning
The core of deep metric learning (DML) involves learning visual similarities in high-dimensional embedding space. One of the main challenges is to generalize from seen classes of training data to unseen classes of test data. Recent works have focused on exploiting past embeddings to increase the num...
Main Authors: | , , |
---|---|
Format: | Article in Journal/Newspaper |
Language: | unknown |
Published: |
arXiv
2021
|
Subjects: | |
Online Access: | https://dx.doi.org/10.48550/arxiv.2103.16940 https://arxiv.org/abs/2103.16940 |
id |
ftdatacite:10.48550/arxiv.2103.16940 |
---|---|
record_format |
openpolar |
spelling |
ftdatacite:10.48550/arxiv.2103.16940 2023-05-15T16:01:31+02:00 Learning with Memory-based Virtual Classes for Deep Metric Learning Ko, Byungsoo Gu, Geonmo Kim, Han-Gyu 2021 https://dx.doi.org/10.48550/arxiv.2103.16940 https://arxiv.org/abs/2103.16940 unknown arXiv arXiv.org perpetual, non-exclusive license http://arxiv.org/licenses/nonexclusive-distrib/1.0/ Computer Vision and Pattern Recognition cs.CV Information Retrieval cs.IR Machine Learning cs.LG FOS Computer and information sciences Article CreativeWork article Preprint 2021 ftdatacite https://doi.org/10.48550/arxiv.2103.16940 2022-03-10T14:51:56Z The core of deep metric learning (DML) involves learning visual similarities in high-dimensional embedding space. One of the main challenges is to generalize from seen classes of training data to unseen classes of test data. Recent works have focused on exploiting past embeddings to increase the number of instances for the seen classes. Such methods achieve performance improvement via augmentation, while the strong focus on seen classes still remains. This can be undesirable for DML, where training and test data exhibit entirely different classes. In this work, we present a novel training strategy for DML called MemVir. Unlike previous works, MemVir memorizes both embedding features and class weights to utilize them as additional virtual classes. The exploitation of virtual classes not only utilizes augmented information for training but also alleviates a strong focus on seen classes for better generalization. Moreover, we embed the idea of curriculum learning by slowly adding virtual classes for a gradual increase in learning difficulty, which improves the learning stability as well as the final performance. MemVir can be easily applied to many existing loss functions without any modification. Extensive experimental results on famous benchmarks demonstrate the superiority of MemVir over state-of-the-art competitors. Code of MemVir is publicly available. : Accepted by ICCV2021 Article in Journal/Newspaper 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 Information Retrieval cs.IR Machine Learning cs.LG FOS Computer and information sciences |
spellingShingle |
Computer Vision and Pattern Recognition cs.CV Information Retrieval cs.IR Machine Learning cs.LG FOS Computer and information sciences Ko, Byungsoo Gu, Geonmo Kim, Han-Gyu Learning with Memory-based Virtual Classes for Deep Metric Learning |
topic_facet |
Computer Vision and Pattern Recognition cs.CV Information Retrieval cs.IR Machine Learning cs.LG FOS Computer and information sciences |
description |
The core of deep metric learning (DML) involves learning visual similarities in high-dimensional embedding space. One of the main challenges is to generalize from seen classes of training data to unseen classes of test data. Recent works have focused on exploiting past embeddings to increase the number of instances for the seen classes. Such methods achieve performance improvement via augmentation, while the strong focus on seen classes still remains. This can be undesirable for DML, where training and test data exhibit entirely different classes. In this work, we present a novel training strategy for DML called MemVir. Unlike previous works, MemVir memorizes both embedding features and class weights to utilize them as additional virtual classes. The exploitation of virtual classes not only utilizes augmented information for training but also alleviates a strong focus on seen classes for better generalization. Moreover, we embed the idea of curriculum learning by slowly adding virtual classes for a gradual increase in learning difficulty, which improves the learning stability as well as the final performance. MemVir can be easily applied to many existing loss functions without any modification. Extensive experimental results on famous benchmarks demonstrate the superiority of MemVir over state-of-the-art competitors. Code of MemVir is publicly available. : Accepted by ICCV2021 |
format |
Article in Journal/Newspaper |
author |
Ko, Byungsoo Gu, Geonmo Kim, Han-Gyu |
author_facet |
Ko, Byungsoo Gu, Geonmo Kim, Han-Gyu |
author_sort |
Ko, Byungsoo |
title |
Learning with Memory-based Virtual Classes for Deep Metric Learning |
title_short |
Learning with Memory-based Virtual Classes for Deep Metric Learning |
title_full |
Learning with Memory-based Virtual Classes for Deep Metric Learning |
title_fullStr |
Learning with Memory-based Virtual Classes for Deep Metric Learning |
title_full_unstemmed |
Learning with Memory-based Virtual Classes for Deep Metric Learning |
title_sort |
learning with memory-based virtual classes for deep metric learning |
publisher |
arXiv |
publishDate |
2021 |
url |
https://dx.doi.org/10.48550/arxiv.2103.16940 https://arxiv.org/abs/2103.16940 |
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.2103.16940 |
_version_ |
1766397342845501440 |