Automated Classification System for Tick-Bite Defect on Leather

Table 2

Classification performance comparison to the state-of-the-arts on the same dataset.

Method	Defective: nondefective	Train: validation: testing split	Method	Accuracy (%)	F1-score (%)
[4]	370 : 1527	60 : 5:35	ANN	80	89
[4]	233 : 699	90 : 0:10	AlexNet	74	84
[6]	503 : 1102	70 : 0: 30	Statistical analysis + SVM	74	97
Ours	503 : 1102	49 : 21: 30	XBoost ANN + HOG	94	94

The bold numbers represent the highest values within the experimental results presented.