Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2504.10877 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917986108964864 |
|---|---|
| author | Gharatappeh, Soheil Sekeh, Salimeh Dhiman, Vikas |
| author_facet | Gharatappeh, Soheil Sekeh, Salimeh Dhiman, Vikas |
| contents | RT-DETRs have shown strong performance across various computer vision tasks but are known to degrade under challenging weather conditions such as fog. In this work, we investigate three novel approaches to enhance RT-DETR robustness in foggy environments: (1) Domain Adaptation via Perceptual Loss, which distills domain-invariant features from a teacher network to a student using perceptual supervision; (2) Weather Adaptive Attention, which augments the attention mechanism with fog-sensitive scaling by introducing an auxiliary foggy image stream; and (3) Weather Fusion Encoder, which integrates a dual-stream encoder architecture that fuses clear and foggy image features via multi-head self and cross-attention. Despite the architectural innovations, none of the proposed methods consistently outperform the baseline RT-DETR. We analyze the limitations and potential causes, offering insights for future research in weather-aware object detection. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_10877 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Weather-Aware Object Detection Transformer for Domain Adaptation Gharatappeh, Soheil Sekeh, Salimeh Dhiman, Vikas Computer Vision and Pattern Recognition RT-DETRs have shown strong performance across various computer vision tasks but are known to degrade under challenging weather conditions such as fog. In this work, we investigate three novel approaches to enhance RT-DETR robustness in foggy environments: (1) Domain Adaptation via Perceptual Loss, which distills domain-invariant features from a teacher network to a student using perceptual supervision; (2) Weather Adaptive Attention, which augments the attention mechanism with fog-sensitive scaling by introducing an auxiliary foggy image stream; and (3) Weather Fusion Encoder, which integrates a dual-stream encoder architecture that fuses clear and foggy image features via multi-head self and cross-attention. Despite the architectural innovations, none of the proposed methods consistently outperform the baseline RT-DETR. We analyze the limitations and potential causes, offering insights for future research in weather-aware object detection. |
| title | Weather-Aware Object Detection Transformer for Domain Adaptation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2504.10877 |