Table 4 Performance metrics for classifiers across four stages, based on statistical features.

From: Enhancing image based classification for crop disease detection using a multiclass SVM approach with kernel comparison

Kernel

Stage

TN

TP

FN

FP

Accuracy

Sensitivity

Specificity

Linear

Antracnose

6523

2561

15

10

0.997

0.994

0.998

Healthy

6630

2461

13

5

0.998

0.994

0.999

Red rust

6033

3051

3

21

0.997

0.999

0.996

Yellow rust

8103

999

6

1

0.999

0.994

0.999

RBF

Antracnose

6509

254

35

24

0.991

0.878

0.996

Healthy

6617

2444

30

18

0.994

0.997

0.987

Red rust

6011

3046

8

44

0.994

0.997

0.992

Yellow rust

8103

991

14

1

0.998

0.986

0.999

Quadratic

Antracnose

6470

2489

87

63

0.983

0.966

0.990

Healthy

6628

2404

70

7

0.991

0.971

0.998

Red rust

5848

3054

0

207

0.977

1

0.965

Yellow rust

8104

885

120

0

0.986

0.880

1

Cubic

Antracnose

6309

2512

64

224

0.968

0.975

0.965

Healthy

6388

2384

90

217

0.966

0.963

0.967

Red rust

5939

2625

429

116

0.940

0.859

0.980

Yellow rust

8075

972

33

29

0.993

0.967

0.996