I have been using Slicer to segment the neurovasculature from MRI datasets of several patients. The focus is on the vessels of the circle of Willis, but I also need to include some of the smaller/darker peripheral vessels in the segmentation. My general approach has been a very manual one where I use a fairly conservative threshold and then use the segment editor tools to add in the missing vessels. As you can imagine, this can be very time consuming.
I am wondering which tools are recommended for a more automated approach for segmenting the vessels missed by the thresholding.
I have tried using the VMTK modules (Vesselness filter and Level Set Segmentation) in slicer v4.10, but I either get poor results or I crash my machine as the algorithms try to allocate my entire RAM. Have these tools been updated to be more resource friendly or is there anyway to estimate the require RAM (I currently have 16 GB on my machine, but an upgrade to 32 GB is possible).
Are the tools in the segment editor better for individually segmenting several branches? I could use the “fill between slices tool” to reduce the number of slices I need to work on, but my understanding is that this does not consider the image data when lofting the volume between the marked slices.
Any advice to help speed up my segmentation workflow would be greatly appreciated.