Skip to main content

Table 1 Accuracy assessments of the automatic leaf-panicle segmentation results

From: Leaf to panicle ratio (LPR): a new physiological trait indicative of source and sink relation in japonica rice based on deep learning

ID GG YG YY
mIoU PA mIoU PA mIoU PA
1 0.861 0.937 0.846 0.912 0.900 0.950
2 0.851 0.945 0.860 0.921 0.899 0.952
3 0.840 0.946 0.830 0.902 0.866 0.933
4 0.849 0.958 0.844 0.915 0.861 0.934
5 0.842 0.959 0.845 0.919 0.854 0.936
6 0.837 0.954 0.846 0.921 0.850 0.940
7 0.833 0.954 0.850 0.927 0.856 0.944
8 0.822 0.948 0.847 0.923 0.837 0.933
9 0.825 0.950 0.847 0.924 0.842 0.936
10 0.881 0.987 0.887 0.959 0.890 0.945
11 0.869 0.984 0.864 0.943 0.874 0.931
12 0.857 0.978 0.855 0.934 0.887 0.942
13 0.846 0.979 0.845 0.925 0.889 0.941
14 0.837 0.978 0.841 0.928 0.881 0.935
15 0.847 0.976 0.851 0.936 0.880 0.938
16 0.851 0.978 0.856 0.938 0.885 0.943
17 0.852 0.978 0.860 0.942 0.872 0.937
18 0.863 0.974 0.863 0.946 0.875 0.938
19 0.825 0.962 0.867 0.948 0.880 0.942
20 0.824 0.961 0.871 0.949 0.873 0.938
21 0.824 0.961 0.871 0.950 0.871 0.938
22 0.831 0.960 0.870 0.949 0.871 0.939
23 0.838 0.959 0.868 0.947 0.874 0.941
24 0.837 0.960 0.876 0.948 0.874 0.942
25 0.836 0.961 0.880 0.948 0.874 0.942
26 0.833 0.962 0.879 0.947 0.874 0.942
27 0.833 0.963 0.877 0.947 0.872 0.940
28 0.832 0.962 0.877 0.947 0.871 0.940
29 0.831 0.962 0.869 0.942 0.870 0.941
30 0.833 0.963 0.867 0.942 0.871 0.942
31 0.889 0.966 0.931 0.974 0.934 0.986
32 0.879 0.968 0.934 0.973 0.953 0.986
33 0.901 0.970 0.934 0.973 0.960 0.985
34 0.914 0.973 0.941 0.975 0.968 0.986
35 0.912 0.972 0.939 0.974 0.976 0.989
Mean 0.850 0.964 0.871 0.941 0.970 0.987
Min 0.822 0.937 0.830 0.902 0.837 0.931
Max 0.914 0.987 0.941 0.975 0.976 0.989
Std 0.025 0.012 0.014 0.014 0.014 0.005
  1. GG green panicle with green leaf, YG yellow panicle with green leaf, YY yellow panicle with yellow leaf, mIoU mean intersection-over-union, PA pixel accuracy, Min minimum, Max maximum, Std standard error of mean