CREATE 3D HUMAN VISUALIZATIONS WITH wrnch

wrnch delivers a human-centric computer vision engine, tools, partner integrations, a software developer’s kit, and community portals to help creators and developers create 3D human visualizations that can be incorporated into amazing user experiences.

Get started on our Developer Portal

wrnch Engine
CREATE INTERACTIVE EXPERIENCES WITH HUMAN-CENTRIC COMPUTER VISION

wrnch Engine is a human-centric, computer vision server that enables computers to see and understand human motion, shape, and intent. 
 
With the wrnch Engine, developers can build applications that capture human motion from videos, digitize 3D human visualizations, and deliver engaging solutions with real-time, human interaction. An integration for Epic Games’ Unreal Engine is available so you can seamlessly incorporate motion data in Unreal Engine applications.  
 
These visually aware applications can be deployed on PCs, mobile devices, embedded in devices or in the cloud. For developers who want an on-demand solution, you can use the wrnch Engine for Cloud to extract human pose and motion data from any video input. 
 
To provide maximum performance and accessibility in performing human-centric computer vision, the wrnch Engine leverages hardware chips, high-performance software, cameras, and deep learning inference engines. Learn what it takes to get up and running on the wrnch Engine and explore deployment options that fit your needs.

LEARN MORE: TECHNICAL INFO AND DEPLOYMENT OPTIONS   

WATCH WRNCH ENGINE IN ACTION

Discover what’s new across the wrnch platform: What’s New 2.2.0

wrnch CAPTURESTREAM

Use CaptureStream to evaluate wrnch AI

wrnch CaptureStream is a native iOS application that captures human motion from the device’s camera and streams 3D motion data to host applications. It is designed to serve as a reference application for developing on iOS and the wrnch Engine. It allows developers wanting to create computer vision-enabled applications to discover wrnch functionality and validate their solutions before investing in full-production.​

CaptureStream performs markerless motion capture to record the movements of an actor from a phone camera without requiring the person to wear tracking sensors. Human pose estimation is run against the live video feed to identify skeletal joints, poses and movement. Real-time visualizations of the person in motion are displayed on the phone’s screen with an overlay of a 3D avatar.

Download Now
wrnch RECEIVEVISUALIZE

Get ReceiveVisualize, the companion app

From the wrnch Developer Portal, developers can install wrnch ReceiveVisualize. Human pose metadata is streamed in real-time from CaptureStream to ReceiveVisualize over a local network, allowing developers to visualize 3D character motion in near real-time.

The local network streaming capabilities of CaptureStream makes it easy to ingest 3D motion capture data into your application. With wrnch wrXchange, our communication protocol, you can easily discover and configure network connections to the wrnch Engine servers.

ReceiveVisualize is available as sample code so developers can start developing on and creating their complete vision system proof of concept using the wrnch Engine.

Download Now

PARTNER INTEGRATIONS
PARTNER INTEGRATIONS

Integrate with Unreal Engine

Unreal Engine developers can easily incorporate 3D motion data into their games and applications using the wrnch UE4 integration, ReceiveVisualize for Unreal. 

Download our iOS app, wrnch CaptureStream, to capture full-body 3D motion without the need for any tracking sensors or suits. The app streams motion data to ReceiveVisualize for Unreal in real time, allowing you to animate 3D character movement directly inside Unreal Engine. You can extract FBX data to create, modify, and test your character animation. 

With the ability to perform markerless motion capture and stream data to Unreal in real time, you can deliver interactive experiences, create immersive virtual worlds, and perform virtual production. Our process saves significant cost and time compared to traditional mo-cap methods and animation workflows. 

ReceiveVisualize is available as sample code so you can prototype and create your own interactive applications that detects and responds to human movement. Visit the Developer Portal for more information.

Download from the Developer Portal
wrnch ENGINE CAPABILITIES

COMPUTER VISION

The wrnch Engine uses AI algorithms and deep learning models to capture motion, perform human pose estimationand visualize human poses in 2D or 3D.

MARKERLESS MOTION CAPTURE

The wrnch Engine supports markerless motion capture to perform monocular human motion tracking from any camera, without requiring people to wear cumbersome, expensive sensors.

HPE
HPE image

HUMAN POSE ESTIMATION

The wrnch Engine detects human objects from video input, tracks skeletal joints to infer poses and motion, segments human instances from backgrounds to delineate human shape, and recognizes gestures to predict human intent.

2D & 3D VISUALIZATION

The wrnch Engine interprets the visual form of a human pose in 2D and 3D.

SOFTWARE DEVELOPER'S KIT

wrnch SOFTWARE
DEVELOPER’S KIT

For developers who want to incorporate 3D human pose and motion data in applications, wrnch offers a Software Developer’s Kit (SDK). The wrnch SDK includes:   

  • Application Programming Interface:  A set of C functions and procedures that you can use to identify the camera from which videos are streamed, call the wrnch Engine for processing, and receive 2D or 3D human pose information within their application.  
  • Sample code and documentation: Code samples and product documentation are provided so you can seamlessly incorporate human motion data into your applications.
Download from the Developer Portal