Inception-LSTM Human Motion Recognition with Channel Attention Mechanism

Table 1

The parameters of each layer of the proposed model.

Order number	Layer position	Size	Number of parameters
1	The convolutional layer, 1_1	[64, 1, 1, 1]	128
2	The convolutional layer, 1_2	[64, 64, 1, 1]	4160
3	The convolutional layer, 2_1	[64, 1, 1, 1]	128
4	Convolutional layer 2_21	[128, 64, 1, 3]	24704
5	Convolutional layer 2_22	[128, 128, 3, 1]	49280
6	The convolutional layer, 3_1	[64, 1, 1, 1]	128
7	Convolutional layer: 3_21	[128, 64, 1, 5]	41088
8	Convolutional layer 3_22	[128, 128, 5, 1]	82048
9	Maximum pooling layer	[64, 1, 3, 3]	0
10	Convolutional layer 4	[64, 1, 1, 1]	4160
11	Channel attention block	[1, 5]	385
12	Feature extraction layer	[9, 384, 1, 1]	3465
13	LSTM layer	[64, 18]	5312
14	Fully connected layer	[12, 64]	780