Railway Fastener Fault Diagnosis Based on Generative Adversarial Network and Residual Network Model

Table 7

Accuracies of different methods.

Method	Mixed data (%)	K (%)	L (%)	M (%)	N (%)	O (%)	Average (%)
LBP + SVM	96.21	90.61	91.86	92.56	93.93	94.01	93.20
HOG + SVM	97.32	92.24	93.55	93.89	95.02	95.78	94.63
VGG16	98.84	93.92	94.43	95.44	97.89	98.97	96.58
GAN + ResNet	99.56	95.87	97.41	98.29	99.03	99.34	98.25