New FogTTA framework boosts Transformer object detection in fog

Researchers have developed FogTTA, a novel online test-time adaptation framework to enhance the performance of Transformer-based object detection models in foggy conditions. This framework, applied to the RF-DETR model, allows for real-time adaptation to foggy environments without needing source data or retraining. FogTTA employs a teacher-student design where the deployed model generates pseudo-labels, which are then refined by a student model learning from augmented target samples using Varifocal loss. Experimental results indicate that FogTTA significantly improves detection accuracy and stability while maintaining real-time performance, outperforming existing methods. AI

IMPACT This research could lead to more reliable autonomous driving systems by improving object detection in adverse weather conditions.

RANK_REASON The cluster contains a research paper detailing a new method for object detection. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• arXiv
• FogTTA
• object detection
• RF-DETR
• transformer
• Varifocal loss

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →