Integrated Mediapipe with a CNN Model for Arabic Sign Language Recognition

Table 4

Results for our approach.

	Precision	Recall	F1 score	Support
Alef	0.98	1.00	0.99	50
Beh	0.95	0.95	0.95	38
Teh	0.98	1.00	0.99	40
Theh	1.00	1.00	1.00	59
Jeem	0.92	1.00	0.96	46
Hah	0.98	0.96	0.97	45
Khah	1.00	0.98	0.99	41
Dal	0.94	0.87	0.91	39
Thal	0.98	0.98	0.98	43
Reh	0.88	0.97	0.93	39
Zain	0.97	0.93	0.95	40
Seen	1.00	1.00	1.00	37
Sheen	0.97	1.00	0.99	38
Sad	1.00	1.00	1.00	28
Dad	0.97	1.00	0.99	37
Tah	0.98	1.00	0.99	43
Zah	1.00	0.97	0.99	38
Ain	0.97	0.95	0.96	38
Ghain	0.98	0.98	0.98	42
Feh	0.92	0.96	0.94	25
Qaf	1.00	0.93	0.96	44
Kaf	0.97	1.00	0.99	33
Lam	0.96	1.00	0.98	25
Meem	0.94	0.97	0.96	34
Noon	0.97	0.90	0.93	39
Heh	0.97	1.00	0.99	37
Waw	0.97	0.93	0.95	40
Yeh	1.00	0.97	0.99	35
Accuracy			0.97	1093
Macro avg	0.97	0.97	0.97	1093
Weighted avg	0.97	0.97	0.97	1093