Skip to main content
. 2023 Aug 10;9(1):101340. doi: 10.1016/j.adro.2023.101340

Figure 3.

Figure 3

The 3-dimensional (3D) dose distribution of the ground truth (manual digitization), 3D convolutional neural network (CNN), and 3D CNN plus attention gates–based automatic digitization of 2 patient examples. The white shadow area indicates the high-risk clinical target volume.