A Model for Qur’anic Sign Language Recognition Based on Deep Learning Algorithms

Table 8

QSLRS-CNN accuracy on the test data after resampled applying SMOTE for 200 epochs.

Class name	Precision	Recall	-score	Number of samples in the test data	Number of samples correctly classified	Accuracy
Alif	1.00	1.00	1.00	390	389	99.74%
Ha	1.00	0.98	0.99	392	385	98.21%
Ra	0.98	1.00	0.99	419	418	99.76%
Sin	0.98	0.98	0.98	434	427	98.39%
Sad	0.97	0.98	0.98	433	424	97.92%
Tah	1.00	0.99	0.99	459	453	98.69%
Ayn	0.97	0.98	0.97	401	392	97.76%
Qaf	0.93	0.98	0.95	443	432	97.52%
Kaf	0.95	0.94	0.94	423	402	95.04%
Lam	0.98	0.93	0.95	483	454	94.00%
Mim	0.96	0.97	0.97	421	409	97.15%
Nun	0.97	0.97	0.97	424	411	96.93%
Haa	0.99	0.98	0.98	399	394	98.75%
Ya	0.99	0.98	0.99	399	396	99.25%