Table 4 Comparison with cutting-edge segmentation methods on the BUSI dataset.
From: A hybrid attention network for accurate breast tumor segmentation in ultrasound images
Method | Parameters (Million) | \(J_{i}\) | \(D_{c}\) | \(S_{n}\) | \(A_{cc}\) | \(P_{r}\) | \(S_{p}\) |
|---|---|---|---|---|---|---|---|
BGRA-GSA57 | 101.34 M | 68.75 | 81.43 | 84.14 | 96.34 | 79.01 | 97.63 |
AAU-Net31 | 29.2 M | 69.26 | 78.18 | 86.06 | - | 81.17 | 99.17 |
MCRNet58 | 26.63 M | 69.94 | 82.31 | 81.65 | 96.78 | - | - |
Swin-unet59 | 27.3 M | 74.16 | 79.45 | 83.16 | 96.55 | - | 97.34 |
Eh-former60 | 184.2 M | 76.37 | 84.6 | 87.74 | - | - | 98.17 |
U-Net15 | 34.51 M | 76.54 | 83.13 | 82.83 | 97.91 | 83.94 | 98.81 |
BGRD-TransUNet61 | 109.65 M | 76.77 | 85.08 | 87.62 | 97.14 | 85.89 | - |
Attention U-Net62 | 8.14 M | 77.89 | 85.96 | 85.80 | 97.85 | 86.65 | 98.54 |
Unet++63 | 9.04 M | 81.09 | 88.11 | 87.29 | 98.57 | 89.53 | 99.18 |
DDRA-Net64 | 5.46 M | 89.23 | 75.32 | 92.32 | - | 95.02 | - |
HA-Net (Proposed) | 15.43 M | 94.75 | 97.28 | 97.15 | 99.74 | 97.42 | 99.84 |
