Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Corrupted segmentation masks in UE5.2 first person map #35

Open
4ndr3aR opened this issue Jan 12, 2024 · 0 comments
Open

Corrupted segmentation masks in UE5.2 first person map #35

4ndr3aR opened this issue Jan 12, 2024 · 0 comments

Comments

@4ndr3aR
Copy link

4ndr3aR commented Jan 12, 2024

Hey there,

pardon me for keep opening issues on probably trivial problems to solve, but this one is really defying my debugging abilities.

The problem is: if I use your CameraCaptureToDisk example (without plugin, I can't find a way to compile it), I can capture RGB images and semantic segmentation masks without any problem.


Conversely, if I create a new project (always without plugin) using the first person template and map, and I try to repeat all the steps:

  1. copy CameraCaptureManager.h/.cpp from the plugin to Prjname/Source/Prjname,
  2. add the dependencies to Prjname.Build.cs to avoid linking errors,
  3. copy CameraCaptureManager_BP.uasset to Content/CameraCaptureToDisk/Blueprints and PP_Segmentation.uasset to Content,
  4. build+cook+run
  5. open the project and add the two CameraCaptureManager objects and the two SceneCapture2D objects (children of a CameraActor object for convenience),
  6. have the two CameraCaptureManagers use the respective SceneCapture2D objects and, only for that concerning the segmentation masks, tick the PNG flag and add PP_Segmentation as postprocess material,
  7. edit the level blueprint and make the capture pipeline activate by pressing a key on the keyboard,

I can still capture RGB images but I capture corrupt semantic segmentation masks.


I've tried every possible fix/workaround that came to my mind, the first of which was obviously disabling the alpha channel fix that I proposed in the pull request (it seems to be irrelevant in this sense, it works fine on both CameraCaptureToDisk and my project). What am I doing wrong? Sounds like an alignment issue. The image seems to be there but it's badly aligned, do I see correctly?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant