Distribution Consistency Loss for Large-Scale Remote Sensing Image Retrieval
Remote sensing images are featured by massiveness, diversity and complexity. These features put forward higher requirements for the speed and accuracy of remote sensing image retrieval. The extraction method plays a key role in retrieving remote sensing images. Deep metric learning (DML) captures th...
Published in: | Remote Sensing |
---|---|
Main Authors: | , , |
Format: | Text |
Language: | English |
Published: |
Multidisciplinary Digital Publishing Institute
2020
|
Subjects: | |
Online Access: | https://doi.org/10.3390/rs12010175 |
id |
ftmdpi:oai:mdpi.com:/2072-4292/12/1/175/ |
---|---|
record_format |
openpolar |
spelling |
ftmdpi:oai:mdpi.com:/2072-4292/12/1/175/ 2023-08-20T04:06:09+02:00 Distribution Consistency Loss for Large-Scale Remote Sensing Image Retrieval Lili Fan Hongwei Zhao Haoyu Zhao agris 2020-01-03 application/pdf https://doi.org/10.3390/rs12010175 EN eng Multidisciplinary Digital Publishing Institute Remote Sensing Image Processing https://dx.doi.org/10.3390/rs12010175 https://creativecommons.org/licenses/by/4.0/ Remote Sensing; Volume 12; Issue 1; Pages: 175 deep metric learning remote sensing image retrieval (RSIR) sample balance loss distribution consistency loss Text 2020 ftmdpi https://doi.org/10.3390/rs12010175 2023-07-31T22:57:50Z Remote sensing images are featured by massiveness, diversity and complexity. These features put forward higher requirements for the speed and accuracy of remote sensing image retrieval. The extraction method plays a key role in retrieving remote sensing images. Deep metric learning (DML) captures the semantic similarity information between data points by learning embedding in vector space. However, due to the uneven distribution of sample data in remote sensing image datasets, the pair-based loss currently used in DML is not suitable. To improve this, we propose a novel distribution consistency loss to solve this problem. First, we define a new way to mine samples by selecting five in-class hard samples and five inter-class hard samples to form an informative set. This method can make the network extract more useful information in a short time. Secondly, in order to avoid inaccurate feature extraction due to sample imbalance, we assign dynamic weight to the positive samples according to the ratio of the number of hard samples and easy samples in the class, and name the loss caused by the positive sample as the sample balance loss. We combine the sample balance of the positive samples with the ranking consistency of the negative samples to form our distribution consistency loss. Finally, we built an end-to-end fine-tuning network suitable for remote sensing image retrieval. We display comprehensive experimental results drawing on three remote sensing image datasets that are publicly available and show that our method achieves the state-of-the-art performance. Text DML MDPI Open Access Publishing Remote Sensing 12 1 175 |
institution |
Open Polar |
collection |
MDPI Open Access Publishing |
op_collection_id |
ftmdpi |
language |
English |
topic |
deep metric learning remote sensing image retrieval (RSIR) sample balance loss distribution consistency loss |
spellingShingle |
deep metric learning remote sensing image retrieval (RSIR) sample balance loss distribution consistency loss Lili Fan Hongwei Zhao Haoyu Zhao Distribution Consistency Loss for Large-Scale Remote Sensing Image Retrieval |
topic_facet |
deep metric learning remote sensing image retrieval (RSIR) sample balance loss distribution consistency loss |
description |
Remote sensing images are featured by massiveness, diversity and complexity. These features put forward higher requirements for the speed and accuracy of remote sensing image retrieval. The extraction method plays a key role in retrieving remote sensing images. Deep metric learning (DML) captures the semantic similarity information between data points by learning embedding in vector space. However, due to the uneven distribution of sample data in remote sensing image datasets, the pair-based loss currently used in DML is not suitable. To improve this, we propose a novel distribution consistency loss to solve this problem. First, we define a new way to mine samples by selecting five in-class hard samples and five inter-class hard samples to form an informative set. This method can make the network extract more useful information in a short time. Secondly, in order to avoid inaccurate feature extraction due to sample imbalance, we assign dynamic weight to the positive samples according to the ratio of the number of hard samples and easy samples in the class, and name the loss caused by the positive sample as the sample balance loss. We combine the sample balance of the positive samples with the ranking consistency of the negative samples to form our distribution consistency loss. Finally, we built an end-to-end fine-tuning network suitable for remote sensing image retrieval. We display comprehensive experimental results drawing on three remote sensing image datasets that are publicly available and show that our method achieves the state-of-the-art performance. |
format |
Text |
author |
Lili Fan Hongwei Zhao Haoyu Zhao |
author_facet |
Lili Fan Hongwei Zhao Haoyu Zhao |
author_sort |
Lili Fan |
title |
Distribution Consistency Loss for Large-Scale Remote Sensing Image Retrieval |
title_short |
Distribution Consistency Loss for Large-Scale Remote Sensing Image Retrieval |
title_full |
Distribution Consistency Loss for Large-Scale Remote Sensing Image Retrieval |
title_fullStr |
Distribution Consistency Loss for Large-Scale Remote Sensing Image Retrieval |
title_full_unstemmed |
Distribution Consistency Loss for Large-Scale Remote Sensing Image Retrieval |
title_sort |
distribution consistency loss for large-scale remote sensing image retrieval |
publisher |
Multidisciplinary Digital Publishing Institute |
publishDate |
2020 |
url |
https://doi.org/10.3390/rs12010175 |
op_coverage |
agris |
genre |
DML |
genre_facet |
DML |
op_source |
Remote Sensing; Volume 12; Issue 1; Pages: 175 |
op_relation |
Remote Sensing Image Processing https://dx.doi.org/10.3390/rs12010175 |
op_rights |
https://creativecommons.org/licenses/by/4.0/ |
op_doi |
https://doi.org/10.3390/rs12010175 |
container_title |
Remote Sensing |
container_volume |
12 |
container_issue |
1 |
container_start_page |
175 |
_version_ |
1774717090651963392 |