Features Overview

This page serves as the hub for all Perception-related documentation. We have broken it up into sections to make it easier to search through. Let us know via the GitHub Issues page if you find outdated content or missing documentation!

Fundamentals

Feature	Description
Perception Camera	Captures RGB images and ground truth from a [Camera](https://docs.unity3d.com/Manual/	Output Endpoint	Currently supported output endpoints are: No Output, Perception endpoint, and Solo endpoint.
SOLO Schema	Schema for annotation, metric, and ground-truth data for the default SOLO endpoint
Labeling	A component that marks a GameObject and its descendants with a set of labels
Label Config	An asset that defines a taxonomy of labels for ground truth generation
Randomization	The Randomization tool set lets you integrate domain randomization principles into your simulation.
FAQ	Frequently Asked Questions about common workflows and issues.
Legacy Perception Schema	Schema for annotation, metric, and ground-truth data for the legacy Perception endpoint

Labeling

Feature	Description
Labeling	A component that marks a GameObject and its descendants with a set of labels
Label Config	An asset that defines a taxonomy of labels for ground truth generation
Bounding Box 2D Labeler	Capture 2D bounding boxes for visible labeled objects.
Hierarchical Bounding Boxes	How to combine bounding boxes of objects with parent-child hierarchical relationships during runtime.
Bounding Box 3D Labeler	Capture 3D bounding boxes for visible labeled objects.
Keypoint Labeler	Record the screen locations of specific points on labeled objects such as keypoints on humans.
Metadata Labeler	Reporting object-level or environment-level metadata information during runtime.

Randomization

Feature	Description
Randomization	The Randomization toolset lets you integrate domain randomization principles into your simulation.
Scenarios	Scenarios control execution flow of your simulation – how many iterations to run the simulation, what randomizers to use, etc.
Randomizers	Randomizers encapsulate specific randomization activities to perform during the lifecycle of a randomized simulation.
Randomizer Tags	RandomizerTags are the primary mechanism by which Randomizers query for a certain subset of GameObjects to randomize within a simulation.
Parameters	Parameters are classes that utilize Samplers to deterministically generate random typed values.
Samplers	Samplers generate bounded random float values by sampling from probability distributions.

Data Generation

Feature	Description
Perception Camera	Captures RGB images and ground truth from a Camera.
Dataset Capture	Ensures sensors are triggered at proper rates and accepts data for the JSON dataset.
Output Endpoint	Currently supported output endpoints are: No Output, Perception endpoint, and Solo endpoint.
Metadata Labeler	Reporting object-level or environment-level metadata information during runtime.