Table 2 Per-category performance of the model generalizability at the \(\text {IOU}_{\text {0.5}}\)
Training set | Testing set | Plant AP | Plant AR100 | Plant F1 | Weed AP | Weed AR100 | Weed F1 |
|---|---|---|---|---|---|---|---|
T15U15 training | UGA2018 testing | 0.981 | 1.000 | 0.991 | NA | NA | NA |
UGA2018 training | UGA2018 testing | 0.983 | 0.992 | 0.987 | NA | NA | NA |
T15U18 training | UGA2015 testing | 0.676 | 0.814 | 0.739 | 0.027 | 0.553 | 0.052 |
UGA2015 training | UGA2015 testing | 0.988 | 0.995 | 0.991 | 0.859 | 1.000 | 0.924 |
U15U18 training | TAMU2015 testing | 0.911 | 0.940 | 0.925 | 0.266 | 0.775 | 0.396 |
TAMU2015 training | TAMU2015 testing | 0.999 | 0.999 | 0.999 | 0.980 | 1.000 | 0.990 |