Skip to main content

Table 7 Mean classification accuracy for descriptors computed on the KTH-TIPS2b dataset

From: Evaluation of noise robustness for local binary pattern descriptors in texture classification

 

KTH-TIPS2b dataset

Noise levels

0

0.0002

0.0006

0.0021

0.0070

0.0234

0.0785

0.2637

0.8859

2.9764

10.0000

LBP

89.3

87.2

86.1

80.4

68.5

53.1

43.7

29.4

23.3

18.9

14.9

ILBP

94.5

94.4

93.6

87.4

74.1

59.2

47.4

38.4

27.7

21.8

17.4

MBP

94.1

84.3

73.6

60.2

48.0

40.4

35.4

28.9

23.2

17.5

14.6

LTP

95.5

94.7

93.2

87.2

71.5

55.4

44.1

32.1

25.3

19.5

16.2

ILTP

96.9

96.4

95.7

90.8

79.4

65.4

50.6

39.2

28.0

22.4

17.6

LQP

94.8

93.6

87.2

70.0

57.2

46.5

38.7

27.9

20.3

18.1

14.7

RLBP

93.8

92.7

90.6

83.7

69.9

53.0

42.4

30.4

23.0

19.3

16.0

FLBP

94.3

94.0

93.1

88.6

74.9

56.9

44.2

29.8

21.9

18.2

15.7

SLBP

94.8

94.5

93.5

89.2

76.6

57.8

44.1

30.1

22.8

18.5

16.2

GF

94.6

94.7

94.1

93.8

90.8

83.9

66.4

41.3

22.3

12.9

10.1

GLCM

76.9

76.2

76.0

73.9

71.7

62.5

54.0

39.5

25.0

18.7

12.7

  1. Max value per noise level is in bold and min value in italics.