Entropy based frame exclusion framework for video transmission over next generation networks

With the growing increase of the video traffic and the increasing expectations of users in terms of the acceptable video quality, achieving the users' Quality of Experience (QoE) while maximising the network resource utilisation to avoid any potential loss of revenue for the ISPs had become a challenge. Traditional Admission Control (AC) algorithms have many limitations in terms of achieving the balance between the perceived QoE and the number of admitted video sessions. This paper proposes a novel framework that exploits video traffic characteristics to present an adaptive admission control technique without compromising the perceived QoE for video traffic. More specifically, we apply an information theoretic tool, namely information entropy, to perform frame selection to the incoming video signals. Experiment results highlight the promise of the studied framework and identify possible future applications.