Mobile interactive region-of-interest video streaming with crowd-driven prefetching

  • Authors:

  • Derek Pang;Sherif Halawa;Ngai-Man Cheung;Bernd Girod

  • Affiliations:

  • Stanford University, Stanford, USA;Stanford University, Stanford, USA;Stanford University, Stanford, USA;Stanford University, Stanford, USA

  • Venue:
  • IMMPD '11 Proceedings of the 2011 international ACM workshop on Interactive multimedia on mobile and portable devices
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Small screen sizes, limited bandwidth, and low computational power often prohibit streaming of high-resolution videos to mobile devices over a wireless network. Recent advances in interactive region-of-interest (IRoI) video streaming technology allow users to interactively control pan/tilt/zoom, while providing bit-rate and complexity savings. In this paper, we present a mobile IRoI video streaming system that delivers high-quality interactive video to smartphones and tablets with multi-touch screens. One of the challenges in IRoI video streaming is to enable low-latency interaction when a user switches between different RoIs. We propose a crowd-driven RoI prediction scheme to prefetch future selected regions. Different from previous approaches that extrapolate past user inputs or perform video semantic analysis, our proposed scheme exploits user viewing statistics collected at the server to make RoI predictions. Our experiments show that a crowd-driven prefetching scheme can substantially reduce average RoI switching delays compared to a system without prefetching.