Complete on-device full body tracking solution. Provides real-time body pose, face, hand, and iris tracking from a single RGB webcam or mobile camera.The core framework is render-pipeline independent. However the materials may need to be upgraded to use the samples with other render pipelines.Tracking4All is the original complete AI-powered real-time human tracking solution that enables high-quality body, face, hand, iris tracking using only a standard RGB webcam or mobile camera. It has been in development since 2024 and used in a variety of projects to offer high quality body tracking features.WEBSITE • YOUTUBE DEMOS • DISCORD🚀 KICKSTART YOUR BODY TRACKING PROJECTTracking4All delivers the core foundation you need to quickly create a full product that uses realtime full body tracking features on consumer devices.--- OFFLINE ON-DEVICE PROCESSING ---Pose tracking with 2D/3D landmarks.Hand and finger tracking with 2D/3D landmarks for each hand.Face tracking with key face points - including 52 ARKit blendshape/shapekey coefficients for animating the face of 3D models.Simultanous pose + hand and pose + hand + face tracking.Advanced smoothing options to smooth tracking to 60fps.Run's on consumer hardware including PC and phones.--- FULLY C# SOURCE AVAILABLE ---Puppeteer any humanoid avatar to have it tracked in real-time via FK.Runtime avatar generation is supported.Person and background segmentation.Full C# code availability to modify and customize as you need.--- INCLUDES DOCUMENTATION AND EXAMPLES ---Verified support for; Windows, Mac, iOS, and Android devices.Swap platforms and back-end providers withuot re-writing your application.Designed for real-world use and includes documentation and examples.Get to building your application faster.Has been used for:Games & interactive applications,Virtual try-on,Motion capture & live streaming,Conventions and fun demos,Customized vtubing,Fitness applications.DEVELOPMENT BY A CANADIAN COMPANY WITH FULL DOCUMENTATION & COMPLIMENTARY SUPPORT.Note: Tracking4All is a body tracking framework/SDK that works independently of the render pipeline. Tested on Unity 2022.3+. Runs fully offline and on-device to respect privacy.Thanks to Third-Party Assets that Make Tracking4All Possible:UniVRM/UniGLTF by ousttrue (MIT LIcense)UniVRM/VRM by VRM Consortium (MIT License)MediaPipeUnityPlugin by homuler (MIT License)OneEuroFilterUnity by DarioMazzanti (MIT License)Features:Complete framework will full C# source availability, example scenes and detailed documentation.Complete architecture and data pipeline to expose tracking data from a variety of backend sources (swappable).Fully customizable menu for tweaking tracking settings at runtime including camera swapping.Example scenes for 3D avatar overlay onto person as well as standalone tracking example scenes.Pose tracking (2D + 3D landmarks & skeletal rotations (3D rotations).Hand tracking (2D + 3D landmarks & skeletal rotations (3D rotations).Holistic lite tracking (simultanous 2D/3D pose and hand tracking).Holistic tracking (simultanous 2D/3D pose, hand, and face tracking).Supports mapping motion (mocap) onto any humanoid 3D (Mecanim-compatible) avatar with reconfigurable puppeteering using FK.Configurable smoothing of data to 60+ FPS.Visibility and presence values for landmarks/limbs.Segmentation accesible via provided shader to differentiate background from person being tracked.Supported platforms/OS:Supports Windows, macOS, Linux, Android, and iOS. Runs efficiently on the CPU with optimized models for mobile platforms.Documentation:Included in the Documentation.pdf in the project asset folder.Requires: video input source (webcam/mobile camera), adaquate lighting for optimal tracking performance.
