A New Transfer Learning Ensemble Model with New Training Methods for Gear Wear Particle Recognition

Table 2

Performance comparison of five models.

	DCNN	VGG19	GoogLeNet	CNNE	LCNNE
Precision	0.860	0.873	0.886	0.897	0.913
Recall	0.845	0.863	0.877	0.891	0.910
Accuracy	95.63%	98.25%	98.94%	99.13%	99.50%
Standard deviation	0.5%	0.38%	0.35%	0.25%	0.13%
Time (s)	0.127	0.112	0.106	0.124	0.115