Table 1.

The internal architecture of R-CNN and ML-LSTM.

Layer	Filter-Size	Kernel-Size	Layer-Parameter
Input		-	-
Convolutional (conv)_1	32	7	10,816
conv_2	32	5	20,544
Add [conv_1, conv_2]		-	-
conv_3	64	3	12,352
Add [conv_2, conv_3]		-	-
Convolutional_4	128	1	4160
Add [conv_3, conv_4]		-	-
LSTM	100	-	66,000
LSTM	100	-	80,400
LSTM	100	-	80,400
FC	128	-	12,928
FC	60	-	7740
Total parameters			295,340