How to filter Semantic annotation?

[jemo-mh]

Habitat-Lab: v0.2.4

Habitat-Sim: v0.2.4

Habitat is under active development, and we advise users to restrict themselves to stable releases. Are you using the latest release versions of Habitat-Lab and Habitat-Sim? Your question may already be addressed in the latest versions. We may also not be able to help with problems in earlier versions because they sometimes lack the more verbose logging needed for debugging.

Master branch contains 'bleeding edge' code and should be used at your own risk.

Docs and Tutorials Did you read the docs? https://aihabitat.org/docs/habitat-sim/ yes Did you check out the tutorials? https://aihabitat.org/tutorial/2020/ yes Perhaps your question is answered there. If not, carry on!

question Questions and Help Hi!

I want to filtering out some classes in HM3D dataset. But edit %Datasetname%.semantic.txt doesn't work

Dataset data/scene_datasets.hm3d/minival/00800-TEEsavR23oF/ there are four files

TEEsavR23oF.basis.glb TEEsavR23oF.basis.navmesh TEEsavR23oF.semantic.glb TEEsavR23oF.semantic.txt And I would like to filter just remain ceiling annotation. When I delete other annotations in TEEsavR23oF.semantic.txt, habitat_sim.Simulator generate empty class and printed like [21:50:57:044332]:[Metadata] AttributesManagerBase.h(357)::buildAttrSrcPathsFromJSONAndLoad : No Glob path result for ./data/scene_datasets/hm3d/train/00017-oEPjPNSPmzL/*.basis.scene_instance.json

And edit annotations to other class in TEEsavR23oF.semantic.txt also doesn't work.

I think It is related in *.basis.scene_instance.json in hm3d_annotated_basis.scene_dataset_config.json but I can't find those .basis.scene_instance.json file locally.

Is there any method to filter annotations?? Thanks for your help in advance !

[aclegg3]

Hey @jemo-mh,

A brief overview of the annotations for HM3D:

• TEEsavR23oF.basis.glb is the RGB render mesh only
• TEEsavR23oF.semantic.glb is the same geometry, but contains semantic texture paints (with unique colors for each instance).
• TEEsavR23oF.semantic.txt maps per-instance colors to semantic classes and regions.

I'm not sure exactly what you are trying to do with this filtering. You could:

1. Post-process the semantic images to remove or change pixels of particular class colors before using them for your application.
2. Change the TEEsavR23oF.semantic.txt to map particular instances to other classes (e.g. "ceiling" to "unknown").