Table 5 Class-wise performance comparison of different models on subcategories A1–A20 and the overall average (ALL).

From: A multi-branch feature enhancement-based detection and hierarchical chaotic encryption fusion method for sensitive targets in remote sensing images

 

RE-DETR

v5n

v6n

v7-tiny

v8n

v9t

v10n

v11n

v12n

v13n

Ours

ALL

77.7

85.3

80.9

87.3

85.7

82.6

85.9

86.1

85.0

80.7

90.0

A1

73.9

82.4

76.1

80.7

78.4

75.4

86.7

81.4

74.1

72.9

88.0

A2

81.3

93.6

91.8

97.1

94.5

90.7

94.0

93.9

93.2

90.8

98.2

A3

86.9

89.8

89.0

95.9

92.9

89.7

93.2

90.0

93.2

88.0

98.2

A4

82.6

88.5

87.0

93.1

91.6

87.9

88.6

90.2

91.1

86.8

95.6

A5

69.8

70.2

66.3

68.8

76.4

67.2

71.4

72.6

73.5

65.8

75.1

A6

91.0

96.4

92.7

96.7

96.6

94.5

95.9

97.5

95.5

89.5

96.0

A7

90.9

95.9

90.3

95.4

96.2

94.7

94.2

96.4

97.2

94.9

96.5

A8

87.1

94.7

92.8

90.0

92.8

94.3

94.1

93.0

94.7

93.4

94.5

A9

76.1

90.0

81.7

91.7

91.6

83.7

87.2

88.8

88.0

80.9

94.8

A10

95.0

97.0

95.4

98.2

97.6

94.9

96.9

97.4

97.6

95.2

98.8

A11

70.8

83.2

75.1

81.6

78.1

79.0

84.7

86.8

83.3

81.8

83.3

A12

75.1

89.0

81.5

90.3

88.0

83.4

89.0

90.0

82.5

74.9

93.2

A13

62.5

76.5

76.4

80.8

79.2

74.3

80.2

79.9

80.6

74.8

82.7

A14

82.9

93.3

90.3

92.9

93.3

92.0

93.3

92.7

92.0

90.9

94.1

A15

39.3

45.7

28.3

58.0

47.7

48.8

45.3

50.6

63.5

49.5

57.6

A16

76.9

92.1

94.1

94.1

93.6

88.1

93.2

94.1

91.7

88.7

93.8

A17

88.1

93.9

94.0

97.1

96.3

92.4

95.2

95.9

93.2

90.7

98.0

A18

72.2

77.4

60.4

82.4

76.1

71.6

77.4

71.9

69.8

62.0

87.0

A19

75.0

77.0

75.4

78.4

76.5

73.9

77.6

79.8

72.7

68.6

85.5

A20

77.0

79.0

79.4

83.6

77.1

76.0

80.6

79.5

72.1

74.1

89.6

  1. Significance value bold.