Facial Recognition of Cattle Based on SK-ResNet

Table 2

Cattle facial recognition results.

Model	FLOPs	Size (MB)	Params	Loss (%)	Accuracy (%)
DenseNet	2.86	26.58	6,960,006	7.91	95.84
GoogleNet	1.5	21.39	5,606,054	7.72	95.95
ResNet-50	4.11	89.72	23,520326	17.01	93.53
SKNet	2.38	50.66	13,281,478	10.27	94.26
SK-ResNet	1.65	22.56	5,913,611	3.38	98.42