I have a task of extracting Brain or removing skull from a brain CTA scans. To train the model first of all i need to label the mask. I am using 3D slicer segment editor for it. The process I am following is as explained in Overview | 3D Slicer segmentation recipes (lassoan.github.io) :
- Threshold: 300 to max
- Margin grow by 1.02mm
- Keep largest island
- And finally extracting largest cavity using wrap solidify
After these steps, still the segmentation is not that good, so I am manullly using paint and eraser tool to refine the segmentation. But the main problem is, in the brain CTA its very frustating for me to distinguish whether the particular region is blood vessel or the contrast or the bone. I have been doing this for about 1 months now but still i couldnt do it properly.
Do someone has some suggestion for me? Is there s perfect window level and width for distinguishing them? Or is there any other methods?
I have tried 40,80 window and here the vessels and bone has almost same contrast and also using 150,450 window, but in 3D slicer i cannot exactly provide the window value.
I would be very happy if someone helped me.