You Only Look Once (YOLO)-based object detectors have shown remarkable accuracy for automated brain tumor detection. In this paper, we develop a novel BGFG-YOLO architecture by incorporating Bi-level Routing Attention (BRA), Generalized feature pyramid networks (GFPN), Forth detecting head, and Generalized-IoU (GIoU) bounding box regression loss into YOLOv8. BGFG-YOLO contains an attention mechanism to focus more on important features, and feature pyramid networks to enrich feature representation by merging high-level semantic features with spatial details. Furthermore, we investigate the effect of different attention mechanisms and feature fusions, detection head architectures on brain tumor detection accuracy. Experimental results show that BGFG-YOLO gives a 3.4% absolute increase of mAP50 compared to YOLOv8x, and achieves state-of-the-art on the brain tumor detection dataset Br35H. The code is available at https://github.com/mkang315/BGFG-YOLO.

翻译：基于You Only Look Once (YOLO)的目标检测器在自动脑肿瘤检测中展现出显著精度。本文通过在YOLOv8中引入双层路由注意力机制(BRA)、广义特征金字塔网络(GFPN)、第四检测头以及广义交并比(GIoU)边界框回归损失，提出了一种新颖的BGFG-YOLO架构。BGFG-YOLO包含注意力机制以聚焦重要特征，并通过融合高层语义特征与空间细节的特征金字塔网络增强特征表示。此外，我们研究了不同注意力机制、特征融合方案及检测头架构对脑肿瘤检测精度的影响。实验结果表明，与YOLOv8x相比，BGFG-YOLO的mAP50绝对值提升3.4%，并在脑肿瘤检测数据集Br35H上达到最优性能。代码开源地址：https://github.com/mkang315/BGFG-YOLO。