Skip to main content

Table 1 Liver segmentation accuracy

From: Automated and robust organ segmentation for 3D-based internal dose calculation

Method Dice coefficient% Tumour% Method Dataset
[36] 96.30 65.70 DL 1
[37] 95.90 50.01 DL 1
[38] 95.57 59.36 DL 1
[39] 94.30 72.00 DL 1
[40] 86.00 Non-DL Internal
Our 93.40 DL 1
  1. The accuracy reported is an average of 5 runs. The LiTS dataset, used for the calculations using the reported method, provides independent masks for the hepatic tumours. In our implementation, we combined the tumour masks and the liver masks to determine the total liver masks
  2. DL, deep learning algorithm; non-DL, other methods