Table 2 Model performance test results based on lung CT dataset.
From: An efficient segment anything model for the segmentation of medical images
Model | IoU | DSC | HD |
|---|---|---|---|
Unetr21 | 80.3 | 87.6 | 12.63 |
nn-UNet19 | 80.8 | 88.9 | 11.41 |
Swin-Unetr22 | 81.2 | 88.1 | 11.52 |
Transunet23 | 79.3 | 87.2 | 13.54 |
Medsegdiff24 | 78.2 | 85.1 | 14.25 |
SAM 1 point8 | 51.5 | 65.8 | 33.58 |
MedSAM 1 point9 | 75.6 | 82.4 | 16.32 |
Ours 1 point | 81.3 | 88.2 | 11.42 |
SAM 2 points8 | 63.4 | 72.3 | 29.54 |
MedSAM 2points9 | 76.1 | 83.2 | 15.56 |
Ours 2 points | 82.1 | 89.3 | 11.35 |
SAM BBox8 | 63.6 | 75.6 | 28.32 |
MesSAM BBox9 | 76.5 | 84.1 | 15.10 |
Ours BBox | 82.4 | 90.1 | 10.21 |
