Classification of Woven Fabric Faulty Images Using Convolution Neural Network

Table 4

Performance evaluation of our proposed model.

Classes	Defects	Accuracy (%)	Precision	Recall	F1 score
1	Hole	93.36	0.80	0.80	0.80
2	Spot	93.02	0.82	0.77	0.80
3	Thread defects	94.35	0.80	0.88	0.84
4	Dark thread	93.69	0.78	0.86	0.82
5	Fload	96.35	0.91	0.86	0.88
6	Switch off	96.01	0.91	0.84	0.87