Sample-prototype optimal transport-based universal domain adaptation for remote sensing image classification
Abstract In recent years, there is a growing interest in domain adaptation for remote sensing image scene classification, particularly in universal domain adaptation, where both source and target domains possess their unique private categories. Existing methods often lack precision on remote sensing...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2024-12-01
|
| Series: | Complex & Intelligent Systems |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s40747-024-01747-y |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832571162078478336 |
|---|---|
| author | Xiaosong Chen Yongbo Yang Dong Liu Shengsheng Wang |
| author_facet | Xiaosong Chen Yongbo Yang Dong Liu Shengsheng Wang |
| author_sort | Xiaosong Chen |
| collection | DOAJ |
| description | Abstract In recent years, there is a growing interest in domain adaptation for remote sensing image scene classification, particularly in universal domain adaptation, where both source and target domains possess their unique private categories. Existing methods often lack precision on remote sensing image datasets due to insufficient prior knowledge between the source and target domains. This study aims to effectively distinguish between common and private classes despite large intra-class sample discrepancies and small inter-class sample discrepancies in remote sensing images. To address these challenges, we propose Sample-Prototype Optimal Transport-Based Universal Domain Adaptation (SPOT). The proposed approach comprises two key components. Firstly, we utilize an unbalanced optimal transport algorithm along with a sample complement mechanism to identify common and private classes based on the optimal transport assignment matrix. Secondly, we leverage the optimal transport algorithm to enhance discriminability among different classes while promoting similarity within the same class. Experimental results demonstrate that SPOT significantly enhances classification accuracy and robustness in universal domain adaptation for remote sensing images, underscoring its efficacy in addressing the identified challenges. |
| format | Article |
| id | doaj-art-4a64441daa14488daab3c915457f6424 |
| institution | Kabale University |
| issn | 2199-4536 2198-6053 |
| language | English |
| publishDate | 2024-12-01 |
| publisher | Springer |
| record_format | Article |
| series | Complex & Intelligent Systems |
| spelling | doaj-art-4a64441daa14488daab3c915457f64242025-02-02T12:49:20ZengSpringerComplex & Intelligent Systems2199-45362198-60532024-12-0111111410.1007/s40747-024-01747-ySample-prototype optimal transport-based universal domain adaptation for remote sensing image classificationXiaosong Chen0Yongbo Yang1Dong Liu2Shengsheng Wang3Hunan Engineering Research Center of Advanced Embedded Computing and Intelligent Medical Systems, Xiangnan UniversityEducation Evaluation Center, Air Force Aviation UniversityHunan Engineering Research Center of Advanced Embedded Computing and Intelligent Medical Systems, Xiangnan UniversityCollege of Computer Science and Technology, Jilin UniversityAbstract In recent years, there is a growing interest in domain adaptation for remote sensing image scene classification, particularly in universal domain adaptation, where both source and target domains possess their unique private categories. Existing methods often lack precision on remote sensing image datasets due to insufficient prior knowledge between the source and target domains. This study aims to effectively distinguish between common and private classes despite large intra-class sample discrepancies and small inter-class sample discrepancies in remote sensing images. To address these challenges, we propose Sample-Prototype Optimal Transport-Based Universal Domain Adaptation (SPOT). The proposed approach comprises two key components. Firstly, we utilize an unbalanced optimal transport algorithm along with a sample complement mechanism to identify common and private classes based on the optimal transport assignment matrix. Secondly, we leverage the optimal transport algorithm to enhance discriminability among different classes while promoting similarity within the same class. Experimental results demonstrate that SPOT significantly enhances classification accuracy and robustness in universal domain adaptation for remote sensing images, underscoring its efficacy in addressing the identified challenges.https://doi.org/10.1007/s40747-024-01747-yuniversal domain adaptationremote sensingscene classificationoptimal transport |
| spellingShingle | Xiaosong Chen Yongbo Yang Dong Liu Shengsheng Wang Sample-prototype optimal transport-based universal domain adaptation for remote sensing image classification Complex & Intelligent Systems universal domain adaptation remote sensing scene classification optimal transport |
| title | Sample-prototype optimal transport-based universal domain adaptation for remote sensing image classification |
| title_full | Sample-prototype optimal transport-based universal domain adaptation for remote sensing image classification |
| title_fullStr | Sample-prototype optimal transport-based universal domain adaptation for remote sensing image classification |
| title_full_unstemmed | Sample-prototype optimal transport-based universal domain adaptation for remote sensing image classification |
| title_short | Sample-prototype optimal transport-based universal domain adaptation for remote sensing image classification |
| title_sort | sample prototype optimal transport based universal domain adaptation for remote sensing image classification |
| topic | universal domain adaptation remote sensing scene classification optimal transport |
| url | https://doi.org/10.1007/s40747-024-01747-y |
| work_keys_str_mv | AT xiaosongchen sampleprototypeoptimaltransportbaseduniversaldomainadaptationforremotesensingimageclassification AT yongboyang sampleprototypeoptimaltransportbaseduniversaldomainadaptationforremotesensingimageclassification AT dongliu sampleprototypeoptimaltransportbaseduniversaldomainadaptationforremotesensingimageclassification AT shengshengwang sampleprototypeoptimaltransportbaseduniversaldomainadaptationforremotesensingimageclassification |