<p>Slim-object detection remains a challenging problem in computer vision because the aspect ratios of slim objects vary significantly depending on the viewpoint. Although extensive research has focused on improving feature extraction, such as through Feature Pyramid–based architectures, these improvements alone are insufficient for objects with extreme aspect ratios. Standard single-shot detectors rely on anchor boxes distributed on fixed grids, where anchor distribution is inherently coupled with feature map resolution, leading to inefficient anchor coverage for slim objects. To address this limitation, this paper proposes a Shape-adaptive Anchor strategy that decouples anchor distribution from feature map resolution by determining anchor strides based on anchor shape rather than feature map dimensions. In addition, a feature map reconstruction module is introduced to align feature extraction with the resulting non-uniform anchor distribution. Experimental results on a custom tie detection dataset demonstrate that the proposed method consistently outperforms standard grid-based anchor approaches, achieving a 2.3% improvement in Average Precision while using significantly fewer anchor boxes. This reduction in anchor count leads to lower computational cost and faster inference speed, indicating that decoupling anchor stride from feature map resolution is a key factor for efficient and accurate slim-object detection.</p>