Describing a view alignment framework in 3D Tele-immersion systems

Abstract

A collaborative 3D Tele-immersion (3DTI) system consists of multiple sensor devices sending streams of data to and from each other. We consider the data streams produced from Microsoft Kinect cameras here. A single camera produces large volumes of data every second. A large bandwidth and low network delay is required to support the streaming of large amounts of data in near real time in order to provide a good quality of experience (QoE), High network delay over the Internet causes a view disparity between the multiple sites and an inter-stream disparity between two streams from the same site. Both these issues hinder the QoE. We address these issues here and provide our solutions for it. We touch upon an interpolation based scheme which uses redundant data from the Kinect sensors to only send minimal amounts of data every second. We introduce a framework to put this interpolation scheme to effective use and to measure and minimize the view and inter-stream disparities with multiple sites using a time vector based solution.