Towards Improved Proxy-Based Deep Metric Learning via Data-Augmented Domain Adaptation
Deep Metric Learning (DML) plays an important role in modern computer vision research, where we learn a distance metric for a set of image representations. Recent DML techniques utilize the proxy to interact with the corresponding image samples in the embedding space. However, existing proxy-based D...
Published in: | Proceedings of the AAAI Conference on Artificial Intelligence |
---|---|
Main Authors: | , , , |
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
Association for the Advancement of Artificial Intelligence
2024
|
Subjects: | |
Online Access: | https://ojs.aaai.org/index.php/AAAI/article/view/29400 https://doi.org/10.1609/aaai.v38i13.29400 |
id |
ftjaaai:oai:ojs.aaai.org:article/29400 |
---|---|
record_format |
openpolar |
spelling |
ftjaaai:oai:ojs.aaai.org:article/29400 2024-04-21T08:01:00+00:00 Towards Improved Proxy-Based Deep Metric Learning via Data-Augmented Domain Adaptation Ren, Li Chen, Chen Wang, Liqiang Hua, Kien 2024-03-24 application/pdf https://ojs.aaai.org/index.php/AAAI/article/view/29400 https://doi.org/10.1609/aaai.v38i13.29400 eng eng Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AAAI/article/view/29400/30645 https://ojs.aaai.org/index.php/AAAI/article/view/29400 doi:10.1609/aaai.v38i13.29400 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence Proceedings of the AAAI Conference on Artificial Intelligence; Vol. 38 No. 13: AAAI-24 Technical Tracks 13; 14811-14819 2374-3468 2159-5399 ML: Representation Learning CV: Image and Video Retrieval info:eu-repo/semantics/article info:eu-repo/semantics/publishedVersion 2024 ftjaaai https://doi.org/10.1609/aaai.v38i13.29400 2024-03-27T16:16:00Z Deep Metric Learning (DML) plays an important role in modern computer vision research, where we learn a distance metric for a set of image representations. Recent DML techniques utilize the proxy to interact with the corresponding image samples in the embedding space. However, existing proxy-based DML methods focus on learning individual proxy-to-sample distance, while the overall distribution of samples and proxies lacks attention. In this paper, we present a novel proxy-based DML framework that focuses on aligning the sample and proxy distributions to improve the efficiency of proxy-based DML losses. Specifically, we propose the Data-Augmented Domain Adaptation (DADA) method to adapt the domain gap between the group of samples and proxies. To the best of our knowledge, we are the first to leverage domain adaptation to boost the performance of proxy-based DML. We show that our method can be easily plugged into existing proxy-based DML losses. Our experiments on benchmarks, including the popular CUB-200-2011, CARS196, Stanford Online Products, and In-Shop Clothes Retrieval, show that our learning algorithm significantly improves the existing proxy losses and achieves superior results compared to the existing methods. The code and Appendix are available at: https://github.com/Noahsark/DADA Article in Journal/Newspaper DML AAAI Publications (Association for the Advancement of Artificial Intelligence) Proceedings of the AAAI Conference on Artificial Intelligence 38 13 14811 14819 |
institution |
Open Polar |
collection |
AAAI Publications (Association for the Advancement of Artificial Intelligence) |
op_collection_id |
ftjaaai |
language |
English |
topic |
ML: Representation Learning CV: Image and Video Retrieval |
spellingShingle |
ML: Representation Learning CV: Image and Video Retrieval Ren, Li Chen, Chen Wang, Liqiang Hua, Kien Towards Improved Proxy-Based Deep Metric Learning via Data-Augmented Domain Adaptation |
topic_facet |
ML: Representation Learning CV: Image and Video Retrieval |
description |
Deep Metric Learning (DML) plays an important role in modern computer vision research, where we learn a distance metric for a set of image representations. Recent DML techniques utilize the proxy to interact with the corresponding image samples in the embedding space. However, existing proxy-based DML methods focus on learning individual proxy-to-sample distance, while the overall distribution of samples and proxies lacks attention. In this paper, we present a novel proxy-based DML framework that focuses on aligning the sample and proxy distributions to improve the efficiency of proxy-based DML losses. Specifically, we propose the Data-Augmented Domain Adaptation (DADA) method to adapt the domain gap between the group of samples and proxies. To the best of our knowledge, we are the first to leverage domain adaptation to boost the performance of proxy-based DML. We show that our method can be easily plugged into existing proxy-based DML losses. Our experiments on benchmarks, including the popular CUB-200-2011, CARS196, Stanford Online Products, and In-Shop Clothes Retrieval, show that our learning algorithm significantly improves the existing proxy losses and achieves superior results compared to the existing methods. The code and Appendix are available at: https://github.com/Noahsark/DADA |
format |
Article in Journal/Newspaper |
author |
Ren, Li Chen, Chen Wang, Liqiang Hua, Kien |
author_facet |
Ren, Li Chen, Chen Wang, Liqiang Hua, Kien |
author_sort |
Ren, Li |
title |
Towards Improved Proxy-Based Deep Metric Learning via Data-Augmented Domain Adaptation |
title_short |
Towards Improved Proxy-Based Deep Metric Learning via Data-Augmented Domain Adaptation |
title_full |
Towards Improved Proxy-Based Deep Metric Learning via Data-Augmented Domain Adaptation |
title_fullStr |
Towards Improved Proxy-Based Deep Metric Learning via Data-Augmented Domain Adaptation |
title_full_unstemmed |
Towards Improved Proxy-Based Deep Metric Learning via Data-Augmented Domain Adaptation |
title_sort |
towards improved proxy-based deep metric learning via data-augmented domain adaptation |
publisher |
Association for the Advancement of Artificial Intelligence |
publishDate |
2024 |
url |
https://ojs.aaai.org/index.php/AAAI/article/view/29400 https://doi.org/10.1609/aaai.v38i13.29400 |
genre |
DML |
genre_facet |
DML |
op_source |
Proceedings of the AAAI Conference on Artificial Intelligence; Vol. 38 No. 13: AAAI-24 Technical Tracks 13; 14811-14819 2374-3468 2159-5399 |
op_relation |
https://ojs.aaai.org/index.php/AAAI/article/view/29400/30645 https://ojs.aaai.org/index.php/AAAI/article/view/29400 doi:10.1609/aaai.v38i13.29400 |
op_rights |
Copyright (c) 2024 Association for the Advancement of Artificial Intelligence |
op_doi |
https://doi.org/10.1609/aaai.v38i13.29400 |
container_title |
Proceedings of the AAAI Conference on Artificial Intelligence |
container_volume |
38 |
container_issue |
13 |
container_start_page |
14811 |
op_container_end_page |
14819 |
_version_ |
1796941372675063808 |