Table 3 Results of comparison experiments on training speed and inference speed for different models
From: DB-DCAFN: dual-branch deformable cross-attention fusion network for bacterial segmentation
Methods | Training speed (s/epoch) | Inference speed (ms/image) |
|---|---|---|
U-Net | 3.79 | 11.87 |
CE-Net | 4.97 | 26.03 |
UCtransNet | 8.94 | 32.61 |
Swin-Unet | 4.04 | 11.77 |
TransUNet | 7.55 | 19.56 |
Ours | 5.93 | 20.36 |