FSDN-DETR: Enhancing Fuzzy Systems Adapter with DeNoising Anchor Boxes for Transfer Learning in Small Object Detection
The advancement of Transformer models in computer vision has rapidly spurred numerous Transformer-based object detection approaches, such as DEtection TRansformer. Although DETR’s self-attention mechanism effectively captures the global context, it struggles with fine-grained detail detection, limit...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Mathematics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-7390/13/2/287 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832588041497083904 |
|---|---|
| author | Zhijie Li Jiahui Zhang Yingjie Zhang Dawei Yan Xing Zhang Marcin Woźniak Wei Dong |
| author_facet | Zhijie Li Jiahui Zhang Yingjie Zhang Dawei Yan Xing Zhang Marcin Woźniak Wei Dong |
| author_sort | Zhijie Li |
| collection | DOAJ |
| description | The advancement of Transformer models in computer vision has rapidly spurred numerous Transformer-based object detection approaches, such as DEtection TRansformer. Although DETR’s self-attention mechanism effectively captures the global context, it struggles with fine-grained detail detection, limiting its efficacy in small object detection where noise can easily obscure or confuse small targets. To address these issues, we propose <b>F</b>uzzy <b>S</b>ystem <b>DN</b>N-<b>DETR</b> involving two key modules: Fuzzy Adapter Transformer Encoder and Fuzzy Denoising Transformer Decoder. The fuzzy Adapter Transformer Encoder utilizes adaptive fuzzy membership functions and rule-based smoothing to preserve critical details, such as edges and textures, while mitigating the loss of fine details in global feature processing. Meanwhile, the Fuzzy Denoising Transformer Decoder effectively reduces noise interference and enhances fine-grained feature capture, eliminating redundant computations in irrelevant regions. This approach achieves a balance between computational efficiency for medium-resolution images and the accuracy required for small object detection. Our architecture also employs adapter modules to reduce re-training costs, and a two-stage fine-tuning strategy adapts fuzzy modules to specific domains before harmonizing the model with task-specific adjustments. Experiments on the COCO and AI-TOD-V2 datasets show that FSDN-DETR achieves an approximately 20% improvement in average precision for very small objects, surpassing state-of-the-art models and demonstrating robustness and reliability for small object detection in complex environments. |
| format | Article |
| id | doaj-art-1680f9f470074217bcff9bc5ae5bb709 |
| institution | Kabale University |
| issn | 2227-7390 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Mathematics |
| spelling | doaj-art-1680f9f470074217bcff9bc5ae5bb7092025-01-24T13:40:03ZengMDPI AGMathematics2227-73902025-01-0113228710.3390/math13020287FSDN-DETR: Enhancing Fuzzy Systems Adapter with DeNoising Anchor Boxes for Transfer Learning in Small Object DetectionZhijie Li0Jiahui Zhang1Yingjie Zhang2Dawei Yan3Xing Zhang4Marcin Woźniak5Wei Dong6College of Information and Control Engineering, Xi’an University of Architecture and Technology, Xi’an 710055, ChinaCollege of Information and Control Engineering, Xi’an University of Architecture and Technology, Xi’an 710055, ChinaCollege of Information and Control Engineering, Xi’an University of Architecture and Technology, Xi’an 710055, ChinaCollege of Information and Control Engineering, Xi’an University of Architecture and Technology, Xi’an 710055, ChinaCollege of Information and Control Engineering, Xi’an University of Architecture and Technology, Xi’an 710055, ChinaInstitute of Mathematics, Silesian University of Technology, Kaszubska 23, 44-100 Gliwice, PolandCollege of Information and Control Engineering, Xi’an University of Architecture and Technology, Xi’an 710055, ChinaThe advancement of Transformer models in computer vision has rapidly spurred numerous Transformer-based object detection approaches, such as DEtection TRansformer. Although DETR’s self-attention mechanism effectively captures the global context, it struggles with fine-grained detail detection, limiting its efficacy in small object detection where noise can easily obscure or confuse small targets. To address these issues, we propose <b>F</b>uzzy <b>S</b>ystem <b>DN</b>N-<b>DETR</b> involving two key modules: Fuzzy Adapter Transformer Encoder and Fuzzy Denoising Transformer Decoder. The fuzzy Adapter Transformer Encoder utilizes adaptive fuzzy membership functions and rule-based smoothing to preserve critical details, such as edges and textures, while mitigating the loss of fine details in global feature processing. Meanwhile, the Fuzzy Denoising Transformer Decoder effectively reduces noise interference and enhances fine-grained feature capture, eliminating redundant computations in irrelevant regions. This approach achieves a balance between computational efficiency for medium-resolution images and the accuracy required for small object detection. Our architecture also employs adapter modules to reduce re-training costs, and a two-stage fine-tuning strategy adapts fuzzy modules to specific domains before harmonizing the model with task-specific adjustments. Experiments on the COCO and AI-TOD-V2 datasets show that FSDN-DETR achieves an approximately 20% improvement in average precision for very small objects, surpassing state-of-the-art models and demonstrating robustness and reliability for small object detection in complex environments.https://www.mdpi.com/2227-7390/13/2/287object detectiontransformertransfer learningDEtection TRansformerfuzzy systemadapter |
| spellingShingle | Zhijie Li Jiahui Zhang Yingjie Zhang Dawei Yan Xing Zhang Marcin Woźniak Wei Dong FSDN-DETR: Enhancing Fuzzy Systems Adapter with DeNoising Anchor Boxes for Transfer Learning in Small Object Detection Mathematics object detection transformer transfer learning DEtection TRansformer fuzzy system adapter |
| title | FSDN-DETR: Enhancing Fuzzy Systems Adapter with DeNoising Anchor Boxes for Transfer Learning in Small Object Detection |
| title_full | FSDN-DETR: Enhancing Fuzzy Systems Adapter with DeNoising Anchor Boxes for Transfer Learning in Small Object Detection |
| title_fullStr | FSDN-DETR: Enhancing Fuzzy Systems Adapter with DeNoising Anchor Boxes for Transfer Learning in Small Object Detection |
| title_full_unstemmed | FSDN-DETR: Enhancing Fuzzy Systems Adapter with DeNoising Anchor Boxes for Transfer Learning in Small Object Detection |
| title_short | FSDN-DETR: Enhancing Fuzzy Systems Adapter with DeNoising Anchor Boxes for Transfer Learning in Small Object Detection |
| title_sort | fsdn detr enhancing fuzzy systems adapter with denoising anchor boxes for transfer learning in small object detection |
| topic | object detection transformer transfer learning DEtection TRansformer fuzzy system adapter |
| url | https://www.mdpi.com/2227-7390/13/2/287 |
| work_keys_str_mv | AT zhijieli fsdndetrenhancingfuzzysystemsadapterwithdenoisinganchorboxesfortransferlearninginsmallobjectdetection AT jiahuizhang fsdndetrenhancingfuzzysystemsadapterwithdenoisinganchorboxesfortransferlearninginsmallobjectdetection AT yingjiezhang fsdndetrenhancingfuzzysystemsadapterwithdenoisinganchorboxesfortransferlearninginsmallobjectdetection AT daweiyan fsdndetrenhancingfuzzysystemsadapterwithdenoisinganchorboxesfortransferlearninginsmallobjectdetection AT xingzhang fsdndetrenhancingfuzzysystemsadapterwithdenoisinganchorboxesfortransferlearninginsmallobjectdetection AT marcinwozniak fsdndetrenhancingfuzzysystemsadapterwithdenoisinganchorboxesfortransferlearninginsmallobjectdetection AT weidong fsdndetrenhancingfuzzysystemsadapterwithdenoisinganchorboxesfortransferlearninginsmallobjectdetection |