Interactive 3D Segmentation of Brain MRI with Differential Watersheds

We approach here the interactive 3D segmentation problem in the context of the image foresting transform (IFT) -- a graph-based tool to develop image processing operators -- by introducing the IFT- algorithm to compute sequences of IFTs in a differential way. We instantiate the IFT- to be a watershed transform and validate it for segmentation of brain MR-images. The new method reduces by a factor of 8.4 the user's waiting time during segmentation compared to the non-differential watershed approach. It also provides a revertible operator that leads to higher efficiency gains than our previous differential approach, the IFT+.

2003