A Deep Learning Approach of Vehicle Multitarget Detection from Traffic Video

Table 1

Network framework.

Net Label Parameter	Network framework
	Learning rate	Convolutional layer	Maximum pool layer + mean pool layer	BN layer	Activation function
YOLO-voc-v1.0	0.001	23	5+0	22	22Leaky+1Linear
YOLO-voc-v1.1	0.001	20	5+0	20	19Leaky+1Linear
YOLO-voc-v1.2	0.0001	23	5+0	22	22Leaky+1Linear
YOLO-voc-v1.3	0.01	23	5+0	22	22Leaky+1Linear
YOLO-voc-v1.4	0.001	20	5+1	20	19 Leaky+1Relu
YOLO-voc-v1.5	0.001	20	5+1	20	19Leaky+1Linear