Skip to main content
. 2017 Jan 4;7:1990. doi: 10.3389/fpls.2016.01990

Table 3.

The accuracy of estimates of the length and width of wheat grains by SeedCounter mobile application and SmartGrain.

Experiment ID MAEa (mm) MAPEa (%) rla rwa
Sam_L1 0.284 7.453 0.936 0.816
Sam_L2 0.296 7.576 0.928 0.824
Sam_L3 0.283 7.339 0.932 0.822
Sam_L4 0.327 8.306 0.923 0.811
Sam_L5 0.398 9.081 0.797 0.770
Sam_L6 0.349 8.437 0.875 0.769
Sony_L1 0.313 8.277 0.933 0.765
Sony_L2 0.310 8.121 0.931 0.767
Sony_L3 0.298 7.787 0.937 0.777
Sony_L4 0.327 8.418 0.920 0.755
Sony_L5 0.301 7.727 0.913 0.749
Sony_L6 0.338 8.546 0.899 0.730
DNS_L1 0.295 7.852 0.943 0.774
DNS_L2 0.296 7.688 0.935 0.777
DNS_L3 0.287 7.730 0.950 0.779
DNS_L4 0.311 8.229 0.940 0.780
DNS_L5 0.351 8.798 0.890 0.672
DNS_L6 0.346 8.264 0.890 0.787
SmartGrain 0.305 6.973 0.948 0.886

aMean absolute error (MAE), mean absolute percentage error (MAPE) averaged over length and width, and Pearson correlation coefficients for length (rl) and width (rw) between actual and estimated values.