Multiverse Knowledge

This module translates a scene graph in USD (Universal Scene Description) format into the knowledge graph in OWL format

Usage

1. Create USD sublayer for semantic tagging

python3 scripts/TBox_to_usd.py 
        --in_owl=</path/to/TBox.owl>    # Input TBox ontology
        --out_usd=</path/to/TBox.usda>  # Output USD file representing TBox

2. Naive semantic tagging

python3 scripts/auto_sem_tag.py 
        --in_usd=</path/to/scene.usda>          # Input scene graph in USD
        --in_TBox_usd=</path/to/in_TBox.usda>   # Input USD file representing TBox
        --out_ABox_usd=</path/to/out_ABox.usda> # Output semantic scene graph in USD

3. Construct the knowledge graph

python3 scripts/usd_to_ABox.py
          --in_usd=</path/to/in_ABox.usda>      # Input semantic scene graph in USD
          --in_owl=</path/to/TBox.owl>          # Input upper level ontology
          --out_owl=</path/to/out_ABox.owl>     # Output knowledge graph of the scene

Examples

Go to ../../build/multiverse/modules/multiverse_knowledge and run ctest. The scene graph from ../../tests/multiverse_knowledge/input will be translated in the output folder in ../../tests/multiverse_knowledge.