Pancreas Segmentation on Pancreas CT 82 patients
87.67Mean DSCOptimally achievable superpixel assignments
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Optimally achievable superpixel assignmentsEvaluation Protocol=Four-fold cross-validation2017.01 | 87.67 | 2.21 | 81.59 | 91.71 | 0.16 | 0.04 | 0.1 | 0.26 | |
| 2D FCN + Recurrent NetworkTrain/Test=62/20, # Folds=4-CV2018.04 | 82.3 | — | — | — | — | — | — | — | |
| Multi-Model 2D FCNTrain/Test=62/20, # Folds=4-CV2018.04 | 82.2 | — | — | — | — | — | — | — | |
| Attention U-Net# Folds=5-CV2018.04 | 81.48 | — | — | — | — | — | — | — | |
| HNN-RFEvaluation Protocol=Four-fold cross-validation2017.01 | 81.27 | 6.27 | 50.69 | 88.96 | 0.42 | 0.31 | 0.14 | 2.26 | |
| Holistically Nested 2D FCNStage=2, Train/Test=62/20, # Folds=4-CV2018.04 | 81.2 | — | — | — | — | — | — | — | |
| HNN_meanmaxEvaluation Protocol=Four-fold cross-validation2017.01 | 81.14 | 7.3 | 44.69 | 89.98 | 0.43 | 0.32 | 0.12 | 1.88 | |
| 2D FCNTrain/Test=62/20, # Folds=4-CV2018.04 | 80.3 | — | — | — | — | — | — | — | |
| Method [28]Evaluation Protocol=Four-fold cross-validation2017.01 | 78.01 | 8.2 | 34.11 | 88.65 | 0.6 | 0.55 | 0.15 | 4.37 | |
| Holistically Nested 2D FCNStage=1, Train/Test=62/20, # Folds=4-CV2018.04 | 76.8 | — | — | — | — | — | — | — | |
| 2D U-NetTrain/Test=66/16, # Folds=5-CV2018.04 | 75.7 | — | — | — | — | — | — | — | |
| Single Model 2D FCNTrain/Test=62/20, # Folds=4-CV2018.04 | 75.7 | — | — | — | — | — | — | — | |
| Method [7]Evaluation Protocol=Four-fold cross-validation2017.01 | 71.42 | 10.11 | 23.99 | 86.29 | 1.53 | 1.6 | 0.2 | 10.32 |