Table 1.

Basic training configuration of mask R-CNN.

SL No.	Arguments	Values
1	Image_Min_Dim	800
2	Image_Max_Dim	3100
3	Learning_Rate	0.001
4	Learning_Momentum	0.9
5	Weight_Decay	0.0001
6	Steps_Per_Epoch	100
7	Detection_Min_Confidence	0.9
8	Rpn_Anchor_Scales	[32, 64, 128, 256, 512]
9	Images_Per_Gpu	1
10	Num_Classes	1+1 (“Spike,” “Background”)
11	Training and Validation Data	2000 and 500 (train-test-split with test size 0.2)
12	Testing Data	Varies
13	Epochs	100