Table 22. Performance of proposed model on the dataset before augmentation with various splitting ratios.

Split ratio (Training:validation:test)	Optmizer	Learning Rate	Training Accuracy (%)	Validation Accuracy (%)	Test Accuracy (%)
60:30:10	Adam	0.001	98.54	97.71	96.83
70:20:10	Adam	0.001	99.46	98.73	98.67
75:15:10	Adam	0.001	99.71	98.28	98.62
80:10:10	Adam	0.001	99.34	98.63	98.51