We developed and compared two different Deep-Learning (DL) based approaches to approximating temperature in subject-specific body models. The first involved use of a 2D U-net to predict temperature throughout the body on a slice-by-slice basis, and the second involved use of a 3D U-net to predict temperature in the 3D body. The 3D approach greatly outperformed the 2D approach, and was very fast.
This abstract and the presentation materials are available to members only; a login is required.