With the help of our valuable community we "discovered" an OpenSIM - serious analytical tool for researchers in the field of Biomechanics.
A simple and fast way to use Nuitrack tracking in OpenSIM - is to record skeleton data in CSV format for further import and analysis.
Just install Nuitrack, activate a Nuitrack AI subscription, and record the skeleton data directly from the Nuitrack application.

P.S. In the coming days we'll additionally publish the conversion script and video tutorial, so you literally don't have to write a single line of code to use Nuitrack for OpenSIM research.

The Nuitrack for Unreal Engine 5.4 has been released!
Get Nuitrack Plugin

Get a 10% discount until the end of July using the promo code "UE2407" (at Nuitrack licensing dashboard https://nuitrack.com/#pricing ).

Also check out our new demo!
This video demonstrates how real-time body tracking could be used in a gaming project based on Unreal Engine 5.
Demo project uses Nuitrack features such as skeleton tracking and face tracking.
In the coming months, this demo project will be made publicly available so you can try it live with any supported depth sensor.

Just as a kind reminder: we are constantly improving the Core - our tracking engines.
Our primary tool - The Failure Case Recorder.
If Nuitrack can't track something:

1. you send us a problematic scenario
2. it's our problem from now on :-) so you just have to relax and wait when official Nuitrack release will include a tracking fix which was designed specifically to resolve your needs

It's just as simple as that: "Before" → "After"

P.S. we support a huge lineup of sensors, so there is always something to improve, but we won't stop until you receive a solution

Kinect v2, being no longer produced (hardware) and maintained (software), is still widely used in rehabilitation systems and other motion analytics applications.

Nuitrack enables step-by-step migration from a hard-to-find Kinect v2 without any quality and performance issue:

1. As a first step you could use Nuitrack SDK with your existing Kinect v2 hardware (no need to replace the whole sensor fleet)
2. Moving to a sensor-agnostic Nuitrack SDK allows you to consider virtually any mainstream depth sensor (from Orbbec, Intel, etc) on the market for your further expansio.

Nuitrack is actively maintained and has a large development community!

Don't miss opportunities due to limitations of unsupported/deprecated 3rd-party components — try Nuitrack for free: https://nuitrack.com/.

Installing the Nuitrack plugin is a breeze through the Epic Games Launcher. Simply navigate to the "Plugins" section, select "Nuitrack", restart UE, and voila! The Nuitrack Plugin is ready for use. No need to tinker with files, create folders, or perform any other tasks.

Dive right into prototyping your application with ease using Blueprints.

We are pleased to announce that Nuitrack now supports Docker and is available from the Docker Hub for Linux x64/arm64 platforms.

You could use Nuitrack image as a starting point for testing, and then proceed to the creation of your own application image based on the provided Dockerfile.

Distributing your app has never been easier. No more problems with different customer environments, library conflicts, etc. just set up your docker image and distribute it.

We are glad to present a new video tutorial on the Nuitrack plugin for UE 5.

This tutorial explores the concept of Skeleton Joints and how to use them to visualize the person 3D pose within a virtual scene.

Please subscribe to our channel so that you don't miss the new tutorials.

The latest version of the Nuitrack Unity package adds new tools for virtual fitting room applications.

What's new:

• Components for clothes customization. Just attach clothes to the mannequin on stage and rotate the limbs. You can test new clothes directly in the editor without connecting a sensor or even playing a scene.
• An example with pre-configured clothing assets. You can use it as a starting point for your fitting room application
• Blender asset (mannequin.blend) with the reference rig/skeleton. Can be used for rigging your own models.
• Dedicated ReadMe with a description of the components and instructions on how to get started

As an engineering team we developed a number of computer vision systems and trained a hundreds of AI models at 3DiVi. They are used in products like NuitrackSDK (skeletal tracking) and FaceSDK
(face detection, face recognition and re-identification)and numerous enterprise projects as well.

Our AI team's experience covers:

• Staff and inventory monitoring systems for retail, offices, etc.
• Driver behavior monitoring
• Human action recognition, like fighting, falling down, sitting, etc.
• Gesture recognition for interactive advertisements / digital signage
• Medical expert systems

and many more...

If you have unresolved needs which could be solved by a computer vision - feel free toleave us a message both through a Nuitrack Enterprise feedback form and adedicated page with our CoreAI custom development offering.

In the recent digest we announced RGB tracking demo for conventional web-cameras.

Now we would like to present the new Nuitrack version which provides 2Dinference/tracking results when no raw 3D/depth information is available (due to certainblack clothing materials or direct sunlight).

We are pleased to announce release with the embeddable version of Nuitrack SDK for Android. Now all the relevant libraries could be integrated into your application.

What you will get as a direct consequence:

• Support for Android 10 and later - compatibility with a more strict file storage policy.
• Easy deployment for end-users - only one app should be installed on your customer's device.
• Total installation size is reduced as well.

No more runtime conflicts caused by incompatible versions of Nuitrack.

We're happy to present our real-time body tracking for Unreal Engine 5. It could be used without writing any code, simply connecting Nodes in Blueprints.

It's perfect for the development of games, fitness apps and other interactive projects with body tracking.

Just click the button below to download and try out Nuitrack for Unreal Engine.
Also don't hesitate to leave us a message at Unreal Engine community.

Nuitrack right from its birth (in 2011) was tightly coupled with depth sensing hardware. At that time depth sensing was the only possible foundation for truly three-dimensional spatial understanding.
Time has changed and we see a growing demand for implementation of a typical use-cases with a commodity RGB-only hardware (to lower costs among other reasons).

Number of open-source projects (like OpenPose) help with that, but being a "cheap/inexpensive" solutions, they are fundamentally lacking:

• consistent 3D bio-mechanical / kinematic model
• seamless migration from 2D data source to 3D with increased accuracy
• ability to employ multiple data source though a seamless fusion to increase both the tracking area and accuracy.

Nuitrack Team used both Depth and RGB data for body tracking in the past, e.g. Nuitrack AI engine relies on RGB-based neural network in addition to 3D kinematic model.
We are now making one more step toward Nuitrack Holistic - body motion tracking over any kind of sensory infrastructure.

We present a demo version of RGB-based tracking engine. You are welcome to participate and challenge it!

If RGB-based tracking is relevant for your project - we appreciate any example of challenging video sequence as a feedback (please message us at nuitrack@3divi.com). In exchange we'll provide a free licenses for commercial version of SDK once it'll be available.

Let's build the future of body motion analytics together!

Starting from 0.36.7 Nuitrack supports simultaneous tracking from multiple sensors.

Currently tracking results for each sensor stream are independent, but this milestone opens up a way to the room-scale tracking with increased area coverage and accuracy.

We are relentlessly working on the Nuitrack Holistic - the next major revision of Nuitrack, which is combining/ fusing tracking results from multiple sensors to eliminate blind spots and increase tracking area.

Please stay tuned to get a firsthand experience.

We are inspired by creative digital art community and expanding Nuitrack as a useful tool for creating interactive performances of the future.

Taking steps in this direction we announce a native plugin for TouchDesigner which brings Nuitrack cross-platform and cross-sensor versatility and performance to the world of interactive audiovisual experiences.

We are now at a open beta-test stage and plan to release TD plugin soon.

Please follow this topic and don't hesitate to fill in a feedback form to try out a beta version of the plugin.

This smart device from Orbbec is the next reincarnation of the first-of-its-kind Persee device, which was brought to life through an Indiegogo campaign.

These devices combines rather powerful ARM SoC, proper set of peripherals/interfaces and branded depth sensing module in a small form-factor. This has a number of applications such as smart home appliances and motion/gesture-based gaming.

Nuitrack for Persee+ has been highly anticipated for about a year and its finally supported with the help of Orbbec engineers.

We see a big opportunities in such a combo, so we plan to extend this support and bring to market our proprietary wireless transmission technology which was used in VicoVR wireless sensor.

We are making next steps into the Digital Art Community and expanding our TouchDesigner plugin with our time-proven fast skeletal tracking.

It's already available for trying out, so please feel free to download a package.

Also don't hesitate to leave us a message at TouchDesigner community.

Nuitrack now could be used not only as a library, but also as a standalone executable, which provides tracking results through REST API. It allows you to quickly get started or integrate with languages/ environments which are now lacking an official support. Nuitrack Daemon is currently in beta, but we are aiming for 1:1 correspondence with the feature set of the main Nuitrack library.

As always we are open for feature requests, so please feel free to leave a message at our Community.

We are strongly committed to tracking algorithms improvements, so we are happy to announce a new utility to record data sequences which aren't handled well enough for your usage scenario.

Some of you are familiar with our service desk responses like a "please record *.bag / *.oni from your sensor for us to reproduce this issue". Enough is enough !)

Getting serious, reproduction is an important step for resolution, so we decided to streamline this experience step-by-step.

We introduced simple sensor-independent format, which could be used for any sensor, not only *.bag/*.oni - compatible ones.

Another feature of this format is multi-sensor support as we aimed for a serious multi-sensor fusion as a next-gen feature of Nuitrack.

And last, but not least: we wanted to optimize the experience by swapping "record" and "reproduce" steps. Now you don't have to record numerous sequences in a pursuit of elusive tracking issues, but record once only after actual occurrence of the relevant issue.

Effectively Nuitrack now has an internal "dashcam", which is right there when it's needed!

We wanted to bring attention not only to new features, but also to some existing features, which possibly were missed by some developers in our community.

We believe that some of these features could make a significant impact in terms of new applications and use-cases.

One of such features is Python support. We specifically designed a simplified synchronous API to be compatible with REPL. It allows users to quickly get started and develop semi-interactive simple scripts for different kinds of data analysis pipelines/applications or just as a first phase of application prototyping. We also believe it could be useful to quickly integrate gesture recognition with diverse smart home appliances.

Consumer revision of Astra+ and Astra Pro+ are already supported.

Support for Femto(W) and Persee+ is on its way, so please stay tuned.

Nuitrack now understands whether this person is entering the sensor's field of view for the first time or visited this location before.

Effectively, skeletal engine is tracking persons not only between entering and leaving, but at the time beyond that period.

Either you want to remember the best score for a specific person (during Kinect-like gaming experience) or need to assign a specific set of exercises for a particular patient of rehabilitation / fitness application - you're now all set up!

We are glad to announce the new version of Nuitrack middleware!

Nuitrack got two updates: version 0.35.7 and 0.35.8.

• The new Nuitrack version 0.35.7 is released with the addition of support for the Intel Realsense sensors for the Linux ARM64 platforms (Raspberry Pi 4, NVIDIA Jetson TX2).

Read more about v0.35.7 in release notes.

We are glad to announce the new version of Nuitrack middleware!

The new Nuitrack version 0.35.6 is released with the following improvements:

• Added support for the Azure Kinect camera (Windows x86, Windows x86_64, and Ubuntu amd64). Very lightweight and fast skeletal tracking for Azure Kinect. Nuitrack Pro doesn't require a powerful GPU!
• Added support for the Orbbec Astra Plus camera (Windows x86, Windows x86_64, Ubuntu amd64 and Linux arm32)

Use Nuitrack with Vitruvius for:

• Avatar animation
• Body pose analysis (angles and measurements)
• Basic motion capturing
• Green screen (background removal)

We are glad to announce the new version of Nuitrack middleware!

The new Nuitrack version 0.34.1 is released with the following improvements:

• Added support for the Intel RealSense D455 and L515 cameras
• Fixed issues for NuitrackSDK.unitypackage

Read more in release notes.

We are glad to announce the new version of Nuitrack middleware!

The new Nuitrack version 0.32.0 is released with the following improvements:

• Nuitrack Device API has been implemented. This API allows you to:
  - numerate and activate compatible depth sensors,
  - run Nuitrack with a selected depth sensor,
  - Activation API. Now you can embed Nuitrack into your app (Windows / Linux) and activate it remotely.

Please refer to C++ or C# examples for more details.

• New function that you can use to get Nuitrack version: nuitrack::getVersion() ( nuitrack.GetVersion() for C#).
• Support for Orbbec Astra Stereo S U3.

Read more in release notes.

We are glad to announce the new version of Nuitrack middleware!

The new Nuitrack version 0.31.0 is released with the following improvements:

• Nuitrack_arm64.apk is released (supports Android ABI arm64-v8a)
• Unity 3D: IL2CPP scripting backend support is added
• Stream registration options are supported for Realsense D400 cameras
• Significant optimization of GC usage in nuitrack.net.dll
• Nuitrack activation tool now supports a license activation via proxy
• nuitrack::TerminateException is thrown after Realsense D400 camera is disconnected

Read more in release notes.

Dear customers,

We are updating Nuitrack Online license and making it more flexible. You can now subscribe on a monthly basis with the more convenient automatic renewals (unless you unsubscribe) at the end of the subscribtion term.

Dear customers,

We would like to support our clients and give the opportunity to develop new ideas and businesses in this difficult time.

Get 50% OFF on licenses when you purchase on nuitrack.com before the end of May 2020.

We are glad to announce the new version of Nuitrack middleware!

The new Nuitrack version 0.30.0 is released with the following improvements:

• Added support for Astra Embedded S / Stereo S (Windows/Linux)
• Fixed error with the Mirror settings for Kinect/Kinect V2
• Fixed error with double triggering of onLostuserCallback for UserTracker
• Added HD resolution support of an RGB stream for Orbbec Astra (Windows/Linux/Android)
• Added Structure Sensor (Occipital) support for iOS

The new version of Nuitrack SDK v1.5.0 is released with the following improvements:

• added Structure Sensor (Occipital) support for iOS

We are glad to announce that Nuitrack SDK is now available on GitHub.
From now on you can :

• Сlone or download Nuitrack SDK
• Make a pull request
• Give feedback on issues
• Follow the software changes.

https://github.com/3DiVi/nuitrack-sdk

We are glad to announce the new version of Nuitrack middleware!

The new version of Nuitrack 0.29.0 is released with the following improvements:

• Support for Raspberry Pi 3 and Tinkerboard devices
• Fixed crash when used in ROS
• Librealsense2 library version updated to v2.25.0
• Support for updating online licenses with connection through a proxy server

Coming soon:

• Nuitrack for iOS (Structure sensor)
• Nuitrack AI (new generation algorithm based on deep learning)

Get 50% off with the special promo code when purchasing on nuitrack.com.
Promo code: SUMMER2019
Hurry up! The offer is limited and valid until the end of August 2019.

Please note: This offer cannot be combined with any other offers.

Effective on May 19, 2019 new license program will be introduced. It includes the following:

• New type of portable online license (instead of current annual)
• All platforms (Windows, Linux, Android) will be the same price
• License management portal.

ADVANTAGES:

• Online licenses will be fully portable and will not be linked to the sensor, but it will require a periodic Internet connection. That is, you can move them from one sensor to another.
• The absence of binding to the OS will allow you to change not only the sensor, but also the OS in the process of developing a solution.
• The license management portal allows you to manage all the resources you paid for. That is, you will have access to information on which devices your licenses were activated (in the case of perpetual licenses) and on which devices your licenses are now active (in the case of online licenses). You can also generate activation keys for both a single license and for a group of licenses.

NEW PRICING (Effective May, 19th 2019):

• Nuitrack annual subscription (online license) = $27.99*/$39.99 per year
• Nuitrack perpetual license = $69.99*/$99.99 per sensor

* - Partner Price


Old prices are valid until May 19th, 2019.

We are glad to announce the new version of Nuitrack SDK!

The new version of Nuitrack v0.27.0 is released with the following improvements:

• RealSense library updated to v2.17.0
• Added support for RealSense D435i on Windows 10
• Added support for RealSense D400 series depth cameras on Windows 7
• Fixed error when calling the Nuitrack API function "Nuitrack:: connectOnIssuesUpdate"
• Added support for Kinect V1 with non-rooted Android devices

Coming soon:
- Biometric ID (Face Recognition)
- Annual online license (without linking to the sensor)
- Nuitrack AI - New generation algorithms based on deep learning

We are glad to announce the new version of Nuitrack SDK!

Support for RealSense on Android was added in Nuitrack v0.26.0.

• Recommended RealSense D415/D435 firmware version is 5.11.1.0
• Rooted and non-rooted devices can be used
• USB-OTG support is required.
• RealSense D415 and D435 were tested with the following Android devices:
  • Odroid XU3 (Android 4.4.4 Kitkat) (rooted)
  • Samsung Galaxy S4 (Android 5 Lollipop) (non-rooted)
  • Samsung Galaxy S5 (Android 6 Marshmallow) (non-rooted)
  • Samsung Galaxy S6 (Android 7 Nougat) (non-rooted)
  • Samsung Galaxy S8/S8+ (Android 8 Oreo) (non-rooted)

Coming soon:
- Biometric ID (Face Recognition)
- Annual online license (without linking to the sensor)
- Nuitrack AI - New generation algorithms based on deep learning

We are glad to announce the new version of Nuitrack SDK!

The new version of Nuitrack 0.25.0 is released with the following improvements:
- Updated RealSense2 library to v2.16.5 for Ubuntu x64 and Windows x86 / x86_64, which supports for the D435i* sensor.
- Introduced Nuitrack Installer for Windows.

Coming soon:
- Biometric ID (Face Recognition).
- Annual online license (without linking to the sensor).

*D435i requires FW version 5.10.13+ and is currently available on Linux only.

We are glad to announce the new version of Nuitrack SDK!

The new version of Nuitrack 0.24.0 is released with the following improvements:
- Support for Kinect V2 sensor added for Ubuntu x64 and Windows x86 / x86_64.

Coming soon:
- Windows Installer
- Biometric ID (Face Recognition)
- Annual online license (without linking to the sensor)

We are glad to announce the new version of Nuitrack SDK.
Apart from skeleton data, now you can detect, track and analyse faces of detected users using RGB input! Data such as gender, age and emotions are now available via API.

The new version of Nuitrack 0.23.3 is released with the following improvements:
- Added first implementation of Instance-based API (Beta) - face data is available via JSON string (Ubuntu amd64 and Windows x86/x86_64)
- RealSense2 library updated to version 2.15.0
- Fixed error during enabled depth-to-color registration for RealSense D435
- Updated drivers for Asus Xtion 2

Coming soon:
- Biometric ID (Face Recognition)
- Annual online license (without linking to the sensor)
- Kinect v2 support

The new version of Nuitrack 0.23.0 is released with the following improvements:
- Added support for the ASUS Xtion 2 depth sensor for Linux x64 and Windows x86 / x64
- Fixed an issue that caused high CPU usage for the Intel Realsense D400 series sensors
- Added support for sensor rotation (90, 180, 270 degrees)

New version of Nuitrack 0.21.0 is published with following improvements:
- Support for Intel RealSense D415 sensor (supported camera firmware version 5.8.15 or higher).
- Tracking of up to 6* skeletons.
- Other improvements and bug fixes.