Skip to main content

Table 1 Classification of transverse cross-section wood images at species, genus and family level. The best results obtained are reported in italics

From: Automated classification of wood transverse cross-section micro-imagery from 77 commercial Central-African timber species

 

FR

Species

Genus

Family

  

NF

SR (\(\pm \) std)

NF

SR (\(\pm \) std)

NF

SR (\(\pm \) std)

LPQ

PCA

150

77.88 (\(\pm \) 0.53)

150

80.14 (\(\pm \) 0.57)

125

83.34 (\(\pm \) 0.57)

 

LDA

77

87.42 (\(\pm \) 0.59)

77

88.79 (\(\pm \) 0.59)

77

90.46 (\(\pm \) 0.40)

LBP\(_{12,1}\)

PCA

114

68.50 (\(\pm \) 1.03)

125

70.77 (\(\pm \) 0.89)

125

75.16 (\(\pm \) 0.98)

 

LDA

77

85.02 (\(\pm \) 0.67)

77

86.88 (\(\pm \) 0.64)

77

88.85 (\(\pm \) 0.69)

LBP1\(_{12,2}\)

PCA

107

63.10 (\(\pm \) 0.59)

46

65.59 (\(\pm \) 0.55)

46

70.25 (\(\pm \) 0.56)

 

LDA

77

79.93 (\(\pm \) 0.54)

77

82.24 (\(\pm \) 0.43)

77

84.19 (\(\pm \) 0.35)

LBP2\(_{16,1}\)

PCA

17

54.01 (\(\pm \) 0.66)

17

56.86 (\(\pm \) 0.64)

17

62.26 (\(\pm \) 0.65)

 

LDA

77

72.49 (\(\pm \) 0.62)

77

74.37 (\(\pm \) 0.62)

77

77.46 (\(\pm \) 0.51)

  1. The average success rates and corresponding standard deviations were obtained over ten runs. FR feature reduction method, NF number of features after feature reduction using PCA or LDA, SR success rate, the relative number of correctly classified samples. The best results obtained are reported in italics