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

Table 10

Comparison of the results obtained by the proposed model and other previous models.

Methods		Accuracy		No. of samples	Training time (min)
		Training (%)	Testing (%)
CNN		98.05	97.13	24,137	20.16
Latif et al. [10]		97.6	97.1	50,000	486.0
Alani and Cosma [12]		98.80	97.29	54,049	141.9
Proposed model (100 epochs)	QSLRS-CNN	98.05	97.13	24,137 at training	20.61
	RMO	98.37	97.36		25.1
	RMU	98.66	97.52		21.18
	SMOTE	98.31	97.67		30.56
Proposed model (200 epochs)	QSLRS-CNN	98.75	97.31	24,137 at training	35.410672
	RMO	99.07	97.47		41.070158
	RMU	99.14	97.58		33.452471
	SMOTE	99.54	97.79		42.404353