DFIR-DETR: Frequency-Domain Iterative Refinement and Dynamic Feature Aggregation for Small Object Detection
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914591875792896 |
|---|---|
| author | Gao, Bo Tong, Jingcheng Chen, Xingsheng Yu, Han Li, Zichen |
| author_facet | Gao, Bo Tong, Jingcheng Chen, Xingsheng Yu, Han Li, Zichen |
| contents | Small object detection in complex scenes exposes a fundamental tension in neural network design: backbone attention distributes computation uniformly regardless of content, pyramid necks inflate activation magnitudes during upsampling without norm compensation, and bottleneck convolutions progressively smooth high-frequency edge components through accumulated spatial filtering. In response, we develop DFIR-DETR by tracing each proposed module back to a specific, measurable deficiency in the RT-DETR baseline: uniform attention that ignores spatial complexity, norm drift that destabilises upsampled features, and spatial convolutions that progressively suppress the high-frequency components small objects depend on. On NEU-DET and VisDrone, DFIR-DETR achieves 92.9% and 51.6% mAP50 with only 11.7M parameters and 47.2 GFLOPs, demonstrating consistent gains across two qualitatively different detection domains. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_07078 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DFIR-DETR: Frequency-Domain Iterative Refinement and Dynamic Feature Aggregation for Small Object Detection Gao, Bo Tong, Jingcheng Chen, Xingsheng Yu, Han Li, Zichen Computer Vision and Pattern Recognition Machine Learning Small object detection in complex scenes exposes a fundamental tension in neural network design: backbone attention distributes computation uniformly regardless of content, pyramid necks inflate activation magnitudes during upsampling without norm compensation, and bottleneck convolutions progressively smooth high-frequency edge components through accumulated spatial filtering. In response, we develop DFIR-DETR by tracing each proposed module back to a specific, measurable deficiency in the RT-DETR baseline: uniform attention that ignores spatial complexity, norm drift that destabilises upsampled features, and spatial convolutions that progressively suppress the high-frequency components small objects depend on. On NEU-DET and VisDrone, DFIR-DETR achieves 92.9% and 51.6% mAP50 with only 11.7M parameters and 47.2 GFLOPs, demonstrating consistent gains across two qualitatively different detection domains. |
| title | DFIR-DETR: Frequency-Domain Iterative Refinement and Dynamic Feature Aggregation for Small Object Detection |
| topic | Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2512.07078 |