com.unity.perceptionis in active development. Its features and API are subject to significant change as development progresses.
Perception Package (Unity Computer Vision)
The Perception package provides a toolkit for generating large-scale datasets for computer vision training and validation. It is focused on a handful of camera-based use cases for now and will ultimately expand to other forms of sensors and machine learning tasks.
Visit the Unity Computer Vision page for more information on our tools and offerings!
Join our mailing list. Sign up now to stay up to date on our latest product feature release, upcoming community events, though-leadership blog posts, and more!
Quick Installation Instructions
Get your local Perception workspace up and running quickly. Recommended for users with prior Unity experience.
Perception Tutorial
Detailed instructions covering all the important steps from installing Unity Editor, to creating your first computer vision data generation project, building a randomized Scene, and generating large-scale synthetic datasets by leveraging the power of Unity Simulation. No prior Unity experience required.
Human Pose Labeling and Randomization Tutorial
Step by step instructions for using the keypoint, pose, and animation randomization tools included in the Perception package. It is recommended that you finish Phase 1 of the Perception Tutorial above before starting this tutorial.
FAQ
Check out our FAQ for a list of common questions, tips, tricks, and some sample code.
Verifying Datasets with Dataset Insights
Introduction to Unity's Dataset Insights – a python package for downloading, parsing and analyzing synthetic datasets.
In-depth documentation on individual components of the package.
| 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 |
| 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. |
| Randomization | The Randomization tool set lets you integrate domain randomization principles into your simulation. |
For setup problems or discussions about leveraging the Perception package in your project, please create a new thread on the Unity Computer Vision forum and make sure to include as much detail as possible. If you run into any other problems with the Perception package or have a specific feature request, please submit a GitHub issue.
For any other questions or feedback, connect directly with the Computer Vision team at [email protected].
SynthDet is an end-to-end solution for training a 2D object detection model using synthetic data.
The Robotics Object Pose Estimation Demo & Tutorial demonstrates pick-and-place with a robot arm in Unity. It includes using ROS with Unity, importing URDF models, collecting labeled training data using the Perception package, and training and deploying a deep learning model.
The repository includes two projects for local development in TestProjects folder, one set up for HDRP and the other for URP.
For closest standards conformity and best experience overall, JetBrains Rider or Visual Studio w/ JetBrains Resharper are suggested. For optimal experience, perform the following additional steps:
- To allow navigating to code in all packages included in your project, in your Unity Editor, navigate to
Edit -> Preferences... -> External Toolsand checkGenerate all .csproj files.
- Projects that use the Perception package on Windows or OS X will have a dependency for Python for Unity added to their manifest, in order for the new Dataset Visualizer tool to work. This tool and Python for Unity are not supported on Linux, therefore this dependency should be removed from the project's manifest file if the project is saved on Windows or OSX and opened on Linux.
If you find this package useful, consider citing it using:
@misc{com.unity.perception2021,
title={Unity {P}erception Package},
author={{Unity Technologies}},
howpublished={\url{https://github.com/Unity-Technologies/com.unity.perception}},
year={2020}
}