multi-scale ship detection

Sharpening ocean intelligence with validated SAR ship detection

Validated SAR ship detection has never been more precise. Our improved YOLOv8 method sharpens ocean intelligence by integrating a CA attention module, DSConv convolution, and Shape-IoU loss. Ablation experiments confirm significant gains in precision, recall, and mAP@0.5. The trade-off is deliberate: we sacrifice some fine localization speed for higher conventional detection accuracy. This model excels under complex background clutter and shows strong cross-domain generalization. For deeper coverage on related detection challenges, see our article "From orbit to insight: unified deep learning sharpens oil spill detection."