Table 3. Testing results of different classes in different indicators.

Class	Precision	Recall	mAP@0.5	F1
all	0.866	0.967	0.962	0.914
mask	0.846	0.971	0.952	0.906
unmask	0.884	0.963	0.972	0.922