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

Table 2

Experimental results of different models on the VOC2007 and VOC2012 testing sets.

Methods	The backbone network	AP(%)
YOLOv1 448	VGG16	66.4
YOLOv2 544	Darknet19	78.6
YOLOv3 416	Darknet53	79.3
YOLOv4 640	CSPDarknet53	88.9
YOLOv5s	CSPDarknet53	85.7
Faster RCNN	ResNet-101	76.4
RefineDet512 [46]	VGfotG-16	80.1
R-FCN-3000 [47]	ResNet-101	80.5
DES512 [48]	VGG-16	80.3
DSSD [49]	ResNet-101	81.5
ASSD [50]	ResNet-101	83.0
Earf-YOLO	CSPDarknet53	91.7