OPTAR: Automatic Coordinate Frame Registration between OpenPTrack and Google ARCore using Ambient Visual Features

This thesis presents a system for the estimation of the coordinate frame registration between OpenPTrack and Google ARCore. OpenPTrack is a multi-camera solution that integrates people tracking, skeleton tracking, and pose recognition. ARCore is a framework for the development of Augmented Reality applications on smartphones. The transformation between the two coordinate frames is obtained by exploiting visual features observed by both the phone and OpenPTrack cameras.