Skip to main content

Table 8 Classification result on each extracted set of features for apple rust disease

From: A probabilistic segmentation and entropy-rank correlation-based feature selection approach for the recognition of fruit diseases

Method

Features

Performance measures

 

SLBP

Harlick

Color

Sen (%)

Spec (%)

Accuracy (%)

DT

  

92.5

92.7

92.6

  

 

75.5

81.8

76.6

   

79.9

85.5

78.9

QSVM

  

93.8

92.7

90.4

  

 

79.9

80.0

79.8

   

87.2

87.3

87.2

CSVM

  

93.8

94.5

93.6

  

 

80.65

81.8

80.9

   

81.2

85.5

81.9

QDA

  

83.25

69.1

80.9

  

 

79.25

76.4

78.7

   

83.8

72.7

81.9

Fine KNN

  

89.5

94.5

90.4

  

 

77.2

80.0

77.7

   

88.5

87.3

88.3

WKNN

  

92.3

87.3

91.5

  

 

85.9

87.3

86.2

   

86.6

83.6

86.2

EBT

  

92.5

92.7

90.4

  

 

85.9

87.3

86.2

   

89.4

89.1

89.4

ESDA

  

92.5

92.7

91.5

  

 

81.5

83.6

81.9

   

88.9

85.5

88.7

Multi-class SVM

  

93.4

94.5

93.6

  

 

87.6

85.5

87.2

   

90.7

89.1

90.4

  1. The bold values indicate best results