Soft Segmented Randomization: Enhancing Domain Generalization in SAR ATR for Synthetic-to-Measured
Synthetic aperture radar technology is crucial for high-resolution imaging under various conditions; however, the acquisition of real-world synthetic aperture radar data for deep learning-based automatic target recognition remains challenging due to high costs and data availability issues. To overco...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10763519/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850221054050435072 |
|---|---|
| author | Minjun Kim Ohtae Jang Haekang Song Heesub Shin Jaewoo Ok Minyoung Back Jaehyuk Youn Sungho Kim |
| author_facet | Minjun Kim Ohtae Jang Haekang Song Heesub Shin Jaewoo Ok Minyoung Back Jaehyuk Youn Sungho Kim |
| author_sort | Minjun Kim |
| collection | DOAJ |
| description | Synthetic aperture radar technology is crucial for high-resolution imaging under various conditions; however, the acquisition of real-world synthetic aperture radar data for deep learning-based automatic target recognition remains challenging due to high costs and data availability issues. To overcome these challenges, synthetic data generated through simulations have been employed, although discrepancies between synthetic and real data—stemming from factors such as background clutter and target signature differences—can degrade model performance. In this study, we introduce a novel framework, soft segmented randomization, designed to reduce domain discrepancy and improve the generalization of synthetic aperture radar automatic target recognition models. The soft segmented randomization framework applies a Gaussian mixture model to segment target and clutter regions softly, introducing randomized variations that align the synthetic data’s statistical properties more closely with those of real-world data. Experimental results demonstrate that the proposed soft segmented randomization framework significantly enhances model performance on measured synthetic aperture radar data, making it a promising approach for robust automatic target recognition in scenarios with limited or no access to measured data. |
| format | Article |
| id | doaj-art-b5893f13c6894bebb35dc29c97d46d5e |
| institution | OA Journals |
| issn | 2169-3536 |
| language | English |
| publishDate | 2024-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-b5893f13c6894bebb35dc29c97d46d5e2025-08-20T02:06:50ZengIEEEIEEE Access2169-35362024-01-011217580117581610.1109/ACCESS.2024.350457410763519Soft Segmented Randomization: Enhancing Domain Generalization in SAR ATR for Synthetic-to-MeasuredMinjun Kim0https://orcid.org/0009-0000-6079-4170Ohtae Jang1Haekang Song2Heesub Shin3Jaewoo Ok4Minyoung Back5Jaehyuk Youn6Sungho Kim7https://orcid.org/0000-0002-5401-2459Department of Electronic Engineering, Yeungnam University, Gyeongsan-si, Gyeongsangbuk-do, South KoreaDepartment of Electrical Engineering, Pohang University of Science and Technology (POSTECH), Pohang, South KoreaDepartment of Electronic Engineering, Yeungnam University, Gyeongsan-si, Gyeongsangbuk-do, South KoreaLIG Nex1 Company Ltd., Yongin, South KoreaLIG Nex1 Company Ltd., Yongin, South KoreaLIG Nex1 Company Ltd., Yongin, South KoreaLIG Nex1 Company Ltd., Yongin, South KoreaDepartment of Electronic Engineering, Yeungnam University, Gyeongsan-si, Gyeongsangbuk-do, South KoreaSynthetic aperture radar technology is crucial for high-resolution imaging under various conditions; however, the acquisition of real-world synthetic aperture radar data for deep learning-based automatic target recognition remains challenging due to high costs and data availability issues. To overcome these challenges, synthetic data generated through simulations have been employed, although discrepancies between synthetic and real data—stemming from factors such as background clutter and target signature differences—can degrade model performance. In this study, we introduce a novel framework, soft segmented randomization, designed to reduce domain discrepancy and improve the generalization of synthetic aperture radar automatic target recognition models. The soft segmented randomization framework applies a Gaussian mixture model to segment target and clutter regions softly, introducing randomized variations that align the synthetic data’s statistical properties more closely with those of real-world data. Experimental results demonstrate that the proposed soft segmented randomization framework significantly enhances model performance on measured synthetic aperture radar data, making it a promising approach for robust automatic target recognition in scenarios with limited or no access to measured data.https://ieeexplore.ieee.org/document/10763519/Gaussian mixture modeldata augmentationdeep learningdomain randomizationdomain generalizationSAR-ATR |
| spellingShingle | Minjun Kim Ohtae Jang Haekang Song Heesub Shin Jaewoo Ok Minyoung Back Jaehyuk Youn Sungho Kim Soft Segmented Randomization: Enhancing Domain Generalization in SAR ATR for Synthetic-to-Measured IEEE Access Gaussian mixture model data augmentation deep learning domain randomization domain generalization SAR-ATR |
| title | Soft Segmented Randomization: Enhancing Domain Generalization in SAR ATR for Synthetic-to-Measured |
| title_full | Soft Segmented Randomization: Enhancing Domain Generalization in SAR ATR for Synthetic-to-Measured |
| title_fullStr | Soft Segmented Randomization: Enhancing Domain Generalization in SAR ATR for Synthetic-to-Measured |
| title_full_unstemmed | Soft Segmented Randomization: Enhancing Domain Generalization in SAR ATR for Synthetic-to-Measured |
| title_short | Soft Segmented Randomization: Enhancing Domain Generalization in SAR ATR for Synthetic-to-Measured |
| title_sort | soft segmented randomization enhancing domain generalization in sar atr for synthetic to measured |
| topic | Gaussian mixture model data augmentation deep learning domain randomization domain generalization SAR-ATR |
| url | https://ieeexplore.ieee.org/document/10763519/ |
| work_keys_str_mv | AT minjunkim softsegmentedrandomizationenhancingdomaingeneralizationinsaratrforsynthetictomeasured AT ohtaejang softsegmentedrandomizationenhancingdomaingeneralizationinsaratrforsynthetictomeasured AT haekangsong softsegmentedrandomizationenhancingdomaingeneralizationinsaratrforsynthetictomeasured AT heesubshin softsegmentedrandomizationenhancingdomaingeneralizationinsaratrforsynthetictomeasured AT jaewoook softsegmentedrandomizationenhancingdomaingeneralizationinsaratrforsynthetictomeasured AT minyoungback softsegmentedrandomizationenhancingdomaingeneralizationinsaratrforsynthetictomeasured AT jaehyukyoun softsegmentedrandomizationenhancingdomaingeneralizationinsaratrforsynthetictomeasured AT sunghokim softsegmentedrandomizationenhancingdomaingeneralizationinsaratrforsynthetictomeasured |