LoRA Enhanced Contrastive Learning with SAS Vision Transformers
Leveraging parameter-efficient fine-tuning to overcome underwater imaging constraints in autonomous naval operations.
Automatic target recognition (ATR) with synthetic aperture sonar (SAS) supports advanced naval capabilities, but deep learning is constrained by scarce target imagery, background clutter, and human-in-the-loop assessment. We adapt DINOv3 Vision Transformer (ViT) models to underwa...
ArXiv CS.AI
Source
Generated insight by OnAirToday AI
Adapting Vision Transformers with LoRA for synthetic aperture sonar addresses the critical challenge of data scarcity in specialized underwater reconnaissance. This approach enables more efficient target recognition by fine-tuning large models on limited datasets, significantly enhancing naval autonomy.
Comments
Keep comments relevant and respectful. No links, advertising, self-promotion, or spam.