Deep-Learning-Based Approach for Iraqi and Malaysian Vehicle License Plate Recognition

Table 12

Recognition results for Malaysian plates (YOLOV2).

	Data volume		LP detection				LP recognition
	# training images	# testing images	Training set		Testing set		Training	Testing
			# of failures	# of wrongly detected	# of failures	# of wrongly detected	Recognition rate	Recognition rate
Fold 1	522	127	7	15	9	9	95.79	85.83
Fold 2	473	107	0	5	5	9	98.94	86.92
Fold 3	486	123	10	4	2	8	97.12	91.87
Fold 4	502	133	1	9	4	9	98.01	90.23
Fold 5	504	133	0	7	4	10	98.61	89.47
Average							97.69	88.86