Video quality adaptation to limit energy usage in mobile systems

Abstract

In mobile systems, it is essential to limit the amount of energy consumption if the energy budget is low. This paper proposes a new scheme that makes use of quality scalability in scalable video coding (SVC) to limit energy usage by taking quality-of-experience (QoE) factors into account. To address the tradeoff between QoE and energy consumption, a new dynamic voltage and frequency scaling (DVFS) technique is proposed in which the fidelity level of each video segment is dynamically selected. The proposed scheme estimates the amount of energy consumption based on the analysis of the computational requirements to decode previous frames, and chooses the fidelity level to achieve the best video quality metric (VQM) value within a given energy budget. It was implemented on a recent commercial smartphone in which a new interface is provided for users to control the energy usage based on video quality. Real measurements show that it makes effective use of the remaining battery while maximizing QoE1.