I'm actually reading a paper about brain segmentation but i don't understand whats is an 'Template image', here a part if the paper:
"We denote T a template image and S its segmentation (obtained manually or semi-manually). We also suppose that T is registered on image I to propagate the segmentation S on I. As the boundaries of the regions of interest in I and T are not obvious everywhere in our images, we propose in this paper a new registration algorithm which strongly constrains the deformations of T with deformations learned by registering the segmentation of T on the segmented training images $I_n$, n ∈ [1, · · · , N]. "
Thanks u