Table 3 Ranking of features based on their contribution to the model’s performance. Values are expressed as mean ± standard deviation.

From: Integrating feature selection and explainable CNN for identification and classification of pests and beneficial insects

Rank

Feature

Weight

Rank

Feature

Weight

1

hog_1757

\(0.030 \pm 0.020\)

16

Hollowness

\(0.015 \pm 0.012\)

2

Compactness

\(0.025 \pm 0.018\)

17

Texture Entropy

\(0.015 \pm 0.011\)

3

hog_611

\(0.021 \pm 0.022\)

18

hog_730

\(0.015 \pm 0.019\)

4

Circularity

\(0.022 \pm 0.017\)

19

hog_1578

\(0.013 \pm 0.008\)

5

hog_1294

\(0.020 \pm 0.008\)

20

hog_504

\(0.013 \pm 0.017\)

6

Aspect Ratio

\(0.019 \pm 0.014\)

21

hog_598

\(0.013 \pm 0.008\)

7

Alr

\(0.019 \pm 0.008\)

22

Entropy

\(0.013 \pm 0.017\)

8

hog_1661

\(0.018 \pm 0.007\)

23

hog_56

\(0.013 \pm 0.013\)

9

Roundness

\(0.018 \pm 0.013\)

24

hog_478

\(0.013 \pm 0.008\)

10

Elongation

\(0.017 \pm 0.012\)

25

hog_206

\(0.013 \pm 0.008\)

11

Std_hist

\(0.016 \pm 0.017\)

26

Texture Contrast

\(0.014 \pm 0.010\)

12

hog_1643

\(0.016 \pm 0.010\)

27

Texture Energy

\(0.012 \pm 0.009\)

13

hog_1266

\(0.016 \pm 0.023\)

28

Texture Correlation

\(0.011 \pm 0.008\)

14

hog_60

\(0.015 \pm 0.012\)

29

Texture Variance

\(0.010 \pm 0.007\)

15

hog_298

\(0.015 \pm 0.007\)