Multitarget Detection in Depth-Perception Traffic Scenarios

Table 1

The structure and parameters of the optimized CNN model.

Layer name	Conv1	Conv2	Conv3	Conv4	Conv5	Conv6	Conv7	Conv8
Input size(c, h, w)	3,1280,1280	96,159,159	256,159,159	256,79,79	384,79,79	384,39,39	4096,40,40	4096,40,40
Output size (c, h, w)	96,318,318	256,159,159	256,159,159	384,79,79	384,79,79	4096,40,40	4096,40,40	256,40,40
Kernel(size, stride, pad)	11,4,0	3,1,1	3,1,1	3,1,1	3,1,1	6,1,3	1,1,0	1,1,0
Pooling(size, stride)	3,2		3,2		3,2
Remarks	LBN layer		LBN layer			Dropout0.5	Dropout0.5