Open-Taxonomy-Cityscapes Profile#
This document describes how to author a USD asset that conforms to the Open-Taxonomy-Cityscapes profile.
Open-Taxonomy-Cityscapes is an example profile that demonstrates labeling assets with the Cityscapes taxonomy
instead of NVIDIA’s Wikidata Q-codes. It pairs the Core feature with one closed-vocabulary taxonomy
feature (CITY.001). See Adding a custom taxonomy to build your own.
Profile definition#
The Open-Taxonomy-Cityscapes profile includes the following feature set (see open_taxonomy_profiles.toml).
Each feature’s requirements and dependencies are defined in the feature specifications.
[Open-Taxonomy-Cityscapes]
"0.1.0" = {features = [
{"FET_000_STANDARD" = {version = "0.1.0"}}, # Core
{"FET_041_STANDARD" = {version = "0.1.0"}}, # Cityscapes Labels
]}
Required USD properties and schemas#
Stage metadata (required)#
defaultPrimmust be set.upAxisandmetersPerUnitmust be set on the stage.
Semantic labels (Cityscapes)#
Apply
SemanticsLabelsAPI:cityscapesto prims you want labeled with this taxonomy.Author
token[] semantics:labels:cityscapesusing Cityscapes class names (CITY.001).Values are matched case-insensitively and tolerate
_/-spacing and display-name/alias forms; a value that matches only after normalization passes with a warning suggesting the canonical name.
def Xform "Object" (
prepend apiSchemas = ["SemanticsLabelsAPI:cityscapes"]
)
{
token[] semantics:labels:cityscapes = ["road", "car"]
}
Validation metadata#
When validation is stamped into the asset, the validator records the result under
customLayerData["SimReady_Metadata"]["validation"]. Search for validated_features to inspect
the dated validation result and the feature IDs and versions that were checked.
The result has this general shape:
customLayerData = {
"SimReady_Metadata" = {
"validation" = {
"validated_features" = {
"<YYYY-MM-DD>" = {
"FET_041_STANDARD" = {
"version" = "0.1.0"
"passed" = true
}
}
}
}
}
}
References#
docs/profiles/open_taxonomy_profiles.tomldocs/features/FET_041_STANDARD.mddocs/capabilities/dataset_taxonomies/capability-dataset_taxonomies.mddocs/guides/adding_a_custom_taxonomy.md