A Flexible Approach to High Performance Visualization Enabled Augmented Reality

Abstract

Commonly registration and tracking within Augmented Reality (AR) applications have been built around computer vision techniques that use limited bold markers, which allow for their orientation to be estimated in real-time. All attempts to implement AR without specific markers have increased the computational requirements and some information about the environment is still needed. In this paper we describe a method that not only provides a flexible platform for supporting AR but also seamlessly deploys High Performance Computing (HPC) resources to deal with the additional computational load, as part of the distributed High Performance Visualization (HPV) pipeline used to render the virtual artifacts. Repeatable feature points are extracted from known views of a real object and then we match the best stored view to the users viewpoint using the matched feature points to estimate the objects pose. We also show how our AR framework can then be used in the real world by presenting a markerless AR interface for Transcranial Magnetic Stimulation (TMS).