Table 3.

Average Recall Rate (ARR), Average Precision Rate (APR), sensitivity (SE), specificity (SP) and F-measure (FM) for the state-of-the-art object detection deep learning methods on the EAC dataset based on LOPO-CV

Method	SE	SP	FM
R-CNN	0.60	0.56	0.59
Fast R-CNN	0.64	0.60	0.63
Faster R-CNN	0.88	0.86	0.87
SSD	0.96	0.92	0.94
Mendel et al. [20]	0.94	0.88	0.91
Sommen et al. [16]	0.86	0.87	0.87

Bold values represent the highest values