Table 4 Post-processing prediction metrics for tumor cells under different models.
From: Improving lung cancer pathological hyperspectral diagnosis through cell-level annotation refinement
Method | Accuracy (%) | Specificity (%) | Sensitivity (%) | Dice (%) |
|---|---|---|---|---|
RF | 90.35(±6.05) | 99.99(±0.00) | 0.03(±0.00) | 0.01(±0.06) |
SVM | 90.35(±6.05) | 99.99(±0.01) | 0.07±(0.01) | 0.03(±0.14) |
2D CNN | 90.83(±5.53) | 98.92(±1.35) | 18.50(±21.40) | 21.94(±19.49) |
3D CNN | 91.08(±5.76) | 99.47(±0.66) | 15.44(±18.65) | 20.12(±20.45) |
Ours | 92.36(±5.02) | 97.49(±2.12) | 51.95(±26.00) | 52.19(±20.05) |
