Skip to main content
. 2022 Jul 14;2022:1765550. doi: 10.1155/2022/1765550

Table 2.

Performance comparison to existing methods.

IoU (%) PA (%) Dice (%)
U-Net 62.4 77.5 76.9
DeepLabv3+ 66.8 79.2 80.1
Ours 68.6 81.4 81.3

The mean segmentation accuracy of three vessels over all test data is shown.