A Weakly Supervised Method for Mud Detection in Ores Based on Deep Active Learning

Table 3

The required samples when the model converges.

	Dog and cat mAP = 73.1%	Ore mAP = 71.8%
YOLO-v3	4400	4400
YOLO-AL (RS)	3100	2650
YOLO-AL (LC)	2350	1950