Lightweight SAR Ship Detection via Contrastive Distillation
Researchers have developed a new knowledge distillation framework called SURGE to create more efficient SAR ship detection models. This framework transfers relational geometry from a larger teacher model to a smaller student model using a contrastive learning objective. Experiments on benchmark datasets show significant improvements in detection accuracy, with the student model even surpassing the teacher's performance in some cases. AI
IMPACT Enables more efficient and accurate ship detection in SAR imagery, potentially for real-time applications.