Robust QoE-aware prediction-based dynamic content adaptation framework applied to slides documents in mobile Web conferencing

Habib Louafi, Stéphane Coulombe, Umesh Chandra
2014 Multimedia tools and applications  
In mobile Web conferencing, enterprise documents are generally adapted into JPEG-based Web pages to be visualized on mobile devices. Dynamically identifying the optimal adapted content is very challenging, as a compromise between high visual quality and small delivery time must be made. In this paper, we propose a robust prediction-based dynamic content adaptation framework for JPEG and XHTML formats that computes near-optimal transcoding parameters dynamically with very few computations. The
more » ... oposed framework is comprised of five methods making different compromises between computational complexity and accuracy. For JPEG, the average deviation from optimality (exhaustive method) is 6% and 3% respectively for two of the proposed methods. For XHTML, the average deviation from optimality is 3% and 1% respectively using the same two methods. Moreover, the methods reached optimality 30% and 59% of the time on the tested documents for JPEG and XHTML respectively, which makes the proposed framework very appealing.
doi:10.1007/s11042-014-2030-2 fatcat:cs57r43ccnd5pb4hl6grmukpgy