Feature: FET_045_STANDARD#
Property |
Value |
|---|---|
Feature Name |
|
Runtime |
|
Proprietary Techs |
|
Latest Version |
|
Description#
This feature validates that semantic labels authored on the
SemanticsLabelsAPI:imagenet_1k instance are real members of the
ImageNet-1K (1000 classes) taxonomy. It is a closed-vocabulary check: values
outside the taxonomy fail; values that match only after case/spacing/alias
normalization pass with a warning that suggests the canonical class name.
It is one of several example taxonomy features that show SimReady labeling is not tied to NVIDIA’s Wikidata Q-code taxonomy. See Adding a custom taxonomy.
Dependency Graph#
This feature has no declared feature dependencies. It judges authored label
values only, so it pairs naturally with
FET_011_STANDARD, which requires that geometry carry a
label at all.
Use Cases#
Products or workflows that consume this feature:
Classification-oriented synthetic data generation
Example profile for authoring assets with a non-NVIDIA taxonomy
Requirements#
Version 0.1.0#
Details
Used in Profiles#
This version is used in the following profiles:
Open-Taxonomy-ImageNet1K (
v0.1.0) - Validates one closed-vocabulary taxonomy at a time.
Feature Dependencies#
None.
Requirement List#
Capability: Dataset Taxonomies
Requirements:
-
IN1K.001| Version0.1.0
-
Pipelines#
Source file type:
.usd,.usda,.usdcAuthor
SemanticsLabelsAPI:imagenet_1kandsemantics:labels:imagenet_1kwith canonical ImageNet-1K class names.
Validation or runtime pipeline:
SimReady validation - verifies
IN1K.001taxonomy membership.
Samples#
None.
Benchmarks#
None.
Adapters#
None.