Skip to main content

Table 1 a,b Summary of statistical indicators used to evaluate the prediction power of the direct X-ray images transformation (XRT) model and the CNN model for inferring the kernel weight and shell weight from 2D X-ray scans

From: X-ray driven peanut trait estimation: computer vision aided agri-system transformation

 

Kernel weight/pod (by XRT method)

Kernel weight/pod (by CNN method)

Shell weight/pod (by XRT method)

Shell weight/pod (by CNN method)

(a) Metrics: calibration

 r

0.93

0.97

0.84

0.94

 R2

0.87

0.95

0.71

0.89

 MSE

0.06

0.03

0.02

0.02

 MAE

0.18

0.14

0.09

0.1

 Slope

1.01

0.75

0.78

0.64

 Intercept

0.16

0.09

0.04

0.05

(b) Metrics: testing

 r

0.97

0.97

0.91

0.96

 R2

0.94

0.94

0.82

0.92

 MSE

0.05

0.07

0.01

0.03

 MAE

0.18

0.21

0.08

0.1

 Slope

1.03

0.73

0.84

0.61

 Intercept

0.15

0.1

0.02

0.05

  1. These are: r (Pearson’s correlation coefficient), R2 (coefficient of determination), MSE (mean squared error), MAE (mean absolute error), slope and intercept of relation between the ground-truth observations (kernel and shell weight) and predictions (by XRT and CNN model). The metrics specific to calibrations set (90% of dataset) is in Metrics: calibration, while the metrics of the test set (testing set) is in Metrics: testing