Fusion Attention Mechanism for Foreground Detection Based on Multiscale U-Net Architecture

Table 3

Complete results of the AMU-Net on CDnet-2014 datasets.

Category	Precision	Recall	Specificity	FNR	FPR	PWC	F-measure
PTZ	0.9907	0.9628	0.9999	0.0372	0.0001	0.0285	0.9759
Bad Weather	0.9897	0.9853	0.9998	0.0147	0.0002	0.0425	0.9875
Baseline	0.9970	0.9903	0.9999	0.0097	0.0001	0.0328	0.9936
Camera Jitter	0.9937	0.9889	0.9997	0.0111	0.0003	0.0679	0.9913
Dynamic Bg	0.9965	0.9864	0.9999	0.0136	0.0001	0.0145	0.9914
Intermitt	0.9957	0.9805	0.9997	0.0195	0.0003	0.1613	0.9879
Low Framerate	0.9150	0.8921	0.9998	0.1079	0.0002	0.0475	0.9030
Night Videos	0.9860	0.9701	0.9997	0.0299	0.0003	0.0934	0.9779
Shadow	0.9922	0.9921	0.9996	0.0079	0.0004	0.0657	0.9921
Thermal	0.9907	0.9854	0.9995	0.0146	0.0005	0.0842	0.9880
Turbulence	0.9876	0.9630	0.9999	0.0370	0.0001	0.0256	0.9751
Overall	0.9850	0.9724	0.9998	0.0276	0.0002	0.0603	0.9785