Table 11.

Classification results of the ResNet50 model for MO images of weld defects.

Defect Types	Number of Images	Train Samples	Valid Samples	Test Samples	Recognition Result	Classification Accuracy/%
Crack	1000	800	100	100	99	99
Pit	1000	800	100	100	98	98
Lack of penetration	1000	800	100	100	100	100
Gas pore	1000	800	100	100	98	98
No defects	1000	800	100	100	100	100
Total	5000	4000	500	500	495	99