[Research] incorporating additional information into the latent space of CNN's?

**Context:**
The input is a 3D MRI image, aswell as a position and orientation both in 3D.

Output is a 3D image

Most likely candidate for architecture is 3D-Unet


In general we are trying to predict how a brain will be stimulated based a magnet producing an electric field. The position and orientation I am talking about is the position and orientation of the magnet. If you are interested this paper works on the same problem:

[doi.org/10.1371/journal.pone.0254588](https://doi.org/10.1371/journal.pone.0254588)


**Question:**
How do …

architecture brain cnn context electric general image information machinelearning mri research space unet