Table 2 Comparison between the F1-score in each sample of dataset A in experiment-1 and experiment-2 using ROI-8 and ROI-9. ROI-8 performs best in KIV-2B samples and ROI-9 performs best in non-KIV-2B samples. The sign “- “ is used for “no calculation possible”
Dataset A experiment 1 (in %) | Dataset A experiment-2 (in %) | |||
|---|---|---|---|---|
ROI-8 | ROI-9 | ROI-8 | ROI-9 | |
Non-KIV-2B | 40.0 | 100.0 | 42.1 | 100.0 |
Non-KIV-2B | 50.0 | 100.0 | 62.5 | 90.9 |
KIV-2B | 93.3 | 52.6 | 96.6 | 35.3 |
KIV-2B | 100.0 | 47.6 | 100.0 | 47.6 |
KIV-2B | 96.6 | 25.0 | 96.6 | 25.0 |
KIV-2B | 78.0 | 28.6 | 80.0 | 28.6 |
KIV-2B | 97.3 | 61.5 | 97.1 | 43.5 |
KIV-2B | 100.0 | - | 100 | - |