Table 2 Results of ablation experiments on the TIB-Net dataset.

From: Improved YOLO for long range detection of small drones

Methods

LCbackbone

MAFR

AWLoss

Purning

mAP@0.5 (%)

mAP@0.95 (%)

Model size (MB)

Parameters (M)

FPS

YOLO11n

    

78.4

31.7

5.20

2.59

90.16

Methods(1)

\(\checkmark\)

   

87.9

36.7

4.73

2.33

93.9

Methods(2)

 

\(\checkmark\)

  

90.8

36.9

3.73

1.79

102.8

Methods(3)

  

\(\checkmark\)

 

81.7

32.4

5.20

2.59

90.16

Methods(4)

\(\checkmark\)

\(\checkmark\)

  

93.9

41.0

3.27

1.52

107.8

Methods(5)

\(\checkmark\)

 

\(\checkmark\)

 

87.4

37.6

4.73

2.32

93.4

Methods(6)

 

\(\checkmark\)

\(\checkmark\)

 

90.4

38.3

3.74

1.79

102.8

Methods(7)

\(\checkmark\)

\(\checkmark\)

\(\checkmark\)

 

93.3

42.2

3.27

1.52

107.8

Methods(8)

\(\checkmark\)

\(\checkmark\)

\(\checkmark\)

\(\checkmark\)

95.7

40.9

2.71

1.23

113.88