Adds isolation in 2d to segmentation file. #76
Conversation
|
It looks like the CI failed due to switching between spacing and tabs for indents. I believe that the rest of the code uses four spaces to denote indentation rather than tabs. Looking at your code, I suspect that that may be the only issue? |
Updated for spacing errors
| @@ -25,13 +25,15 @@ def segmentation_timestep(field_in,features_in,dxy,threshold=3e-3,target='maximu | |||
| method: string | |||
| flag determining the algorithm to use (currently watershedding implemented) | |||
| max_distance: float | |||
| maximum distance from a marker allowed to be classified as belonging to that cell | |||
| maximum distance from a marker allowed to be classified as belonging to that cell | |||
| ISO_dilate: int | |||
There was a problem hiding this comment.
Why is 8 the default here? I'm not opposed to the value, I'm just curious.
| @@ -120,6 +123,33 @@ def segmentation_timestep(field_in,features_in,dxy,threshold=3e-3,target='maximu | |||
| else: | |||
| raise ValueError('unknown method, must be watershed') | |||
|
|
|||
| #create isolation, currently only available for 2d tracking | |||
| if field_in.ndim==2: | |||
There was a problem hiding this comment.
Do we want to make this entirely optional? Is running the isolation here necessary for everyone, or would it be better to have a parameter in place to turn it on/off?
There was a problem hiding this comment.
I guess the question here ultimately is: what's the performance penalty?
|
I've added some comments to the code, but generally I'm happy here. I think we will want to wait to merge this until after #78 , but once that's in and this is updated to match the new style, I'm happy to merge this. |
|
Overall this looks good, but I have a couple of thoughts/requests:
The merge conflict is just due to formatting, so we can fix that easily before merging. |
Merge/split as postprocessing step.
This reverts commit d4cbfa2.
|
I've marked this PR as a draft for now. |
This is just an addition to the main tobac segmentation file. The additional parameter is ISO_dilate, a value by which the 2d feature is dilated in order to find the number of neighbors. 1 Neighbor value indicates the feature (which was dilated) does not have a neighbor. Values greater than 1 are the number of neighbors in addition to the base feature. Non-positive values indicate a feature area was not created for the feature id and thus no valid neighbors.