The aim of this project is to provide full-body tracking capability to vr games using only a phone/webcam.
These are projects we are using directly from our driver.
A 2D pose estimation library based on machine learning, we use this to get the basic 2d positions which we are going to reconstruct into a 3d point afterwards.
A framework for image processing, currently we use it for its video capture abilities.
A framework for exposing VR related hardware and software to games, we use it for exposing the virtual trackers.
These are projects we don't use directly but found useful while creating the driver.
This project is the main backbone for the algorithm that reconstructs the 3d position from the outputs of the network.
This project was used to figure out how to properly create a driver that is exposed to the OpenVR server.