Institut numerique

Abstract

One major challenges has guided this research : How efficiency streaming video in a network where bandwidth is limited and unstable, or equipment responsible for receiving and restore the stream are characterized by limited resources ?

To allow for adaptive streaming of video content streaming, we proposed an approach exploiting the color depth available for the frame of the video. Such information enables, among other things, to maintain visual consistency of video when it is returned. In a second step, we studied the performance of our approach by proposing and implementing an algorithm for fragmentation of an image stackable layers, and merge these layers in the original image.

The availability of data in the streaming systems is fundamentally different from that observed in conventional systems, in that the data consumed are time dependent. So this alternative approach in the vision of the problem led us to explore the possibility for the receiver (consumer) to be able to watch the requested video content without the risk of starvation in the receive buffer. In order to solve this problem, we implemented a mechanism for pre-loading, using a system of double buffer rotating, which significantly reduces the latency of the player (client) of our system.

Finally, we describe the design and implementation of a prototype that demonstrates the feasibility of a system for adaptive streaming over the light environment.

Keywords :

streaming video, limited resources, unstable network, adaptive streaming, dual buffer turning, overlapping images.

Page suivante : Glossaire

Retour au menu : Adaptation de flux vidéo en environnements limités et dynamiques : Proposition d’une solution par encodage vidéo multicouche