Page MenuHomePhabricator

3D region growing may have issues when the seed point is on a pixel with value 0
Closed, ResolvedPublic

Description

As discussed today in the MITK meeting I was having trouble with applying 3d region growing to a dataset of mice lungs. The task was to segment the lumen of the airways. I set the seedpoint in the middle of the lumen (see upper left corner in screenshot) and clicked "run segmentation"

image.png (1×2 px, 447 KB)

Unfortunately, no segmentation was created at all.

As a sanity check: when I instead apply region growing 3d to the bright border of another airvessel everything works as expected:

image.png (548×503 px, 85 KB)

My semi-educated guess is that the voxel value of 0 could be giving issues?

Another observation: If I set the seed point to a pixel close to the border (which has a non-0 value but still low intensity) the region growing seems to work (as long as I don't touch the "define thresholds" slider which defaults to [0,1 38.3]:

image.png (560×494 px, 44 KB)

HOWEVER, as soon as I modify the lower bound of the "define thresholds" slider from 0,1 to 0 it again does not work.

Best,

Fabian

Revisions and Commits

rMITK MITK
Restricted Differential Revision
Restricted Differential Revision
Restricted Differential Revision

Event Timeline

kislinsk triaged this task as High priority.Apr 8 2021, 9:44 AM
kahl moved this task from Segmentation to Cycle on the MITK (v2021.10) board.
kahl added a revision: Restricted Differential Revision.Apr 23 2021, 10:26 AM

Deleted branch from rMITK MITK: bugfix/T28437-RegionGrowing3DSeedPoint0.