Computer Vision Software wrnch Launches AI Pose Estimation Extension for NVIDIA Omniverse
Canadian computer vision / deep learning software engineering company, wrnch announced its collaboration with NVIDIA to deliver AI-powered, human pose estimation capabilities in NVIDIA Omniverse™ Machinima. Now, game creators can more easily incorporate convincing 3D characters in their stories.
What is Wrnch?
wrnch Inc. is a computer vision / deep learning software engineering company based in Montreal, Canada – a world-renowned hub for artificial intelligence and visual computing.
Its flagship product, wrnch Engine, processes digital images from cameras and videos using deep learning models to detect humans in images, recognize skeletons and shapes, understand motion and gesture, and deliver results to that computer applications can respond quickly to what has been seen
The technology includes:
wrnch CaptureStream – a free application that enables creators to use a mobile device’s camera to capture the human motion that they’d like to reproduce in an application. As human performance is captured, the Engine detects humans in the video feed and uses powerful human pose estimation algorithms to track skeletal joints to infer digital human pose and motion. The wrnch Engine outputs 2D and 3D pose metadata in real time.
NVIDIA Omniverse Machinima is a mixer application that offers a rich set of tools and extensions that enable artists, creators and developers to:
- Render realistic scenes using materials, surfaces and textures from NVIDIA MDL or imported from third-party asset libraries
- Create breathtaking visual effects using NVIDIA PhysX and Flow extensions
- Animate a character’s face with only a voice recording using NVIDIA Audio2Face technology
- Incorporate 3D characters that mimic human movement captured from video using the wrnch AI Pose Estimator extension
Output a high-fidelity game using the built-in Omniverse RTX Renderer.
“We’ve been proud to work with NVIDIA to integrate wrnch technology into Omniverse Machinima,” said Paul Kruszewski, CEO of wrnch Inc. “With the democratization of motion capture and AI pose estimation, creators across the world can easily generate compelling 3D content to animate characters in stories that were never possible before.”
AI Pose Estimator – an Omniverse extension that enables creators to detect and connect to the wrnch CaptureStream application running on the local network. As human pose metadata is streamed in real time, the extension translates the wrnch metadata into Pixar Universal Scene Description (USD) format to animate 3D characters in Omniverse applications.
“People using Omniverse will be delighted with our collaboration with wrnch Inc. to bring state-of-the art human pose estimation technology to the 3D collaboration and simulation platform,” said Richard Kerris, general manager of Omniverse at NVIDIA. “The wrnch AI Pose Estimator extension simplifies the process of transforming human motion detected in videos into animated 3D characters in an Omniverse application.”