Earf-YOLO: An Efficient Attention Receptive Field Model for Recognizing Symbols of Zhuang Minority Patterns

Figure 1

An overview of Earf-YOLO. Transformer and SRFB are used to optimize YOLOv4.Then, some advanced techniques such as data augmentation, GIoU Loss, Focal Loss, and Soft-NMS were employed to improve the Earf-YOLO detection efficiency on Zhuang pattern symbol datasets. It is worth noting that the feature map in the figure can obtain three prediction results through the gating axial transformer.