SenseQ: Context-Aware Video Quality Adaptation for Optimal Mobile Video Streaming in Dynamic Environments

Abstract

The growth of the mobile device and video streaming market has led to a significant increase in mobile video streaming as a primary mode of media consumption over the Internet. Mobile video streaming systems operate in dynamic environments and contexts that can change over time, posing challenges for the HVS (human visual system) and the quality of video streaming. To address these challenges in the context of IoT, we introduce SenseQ, a novel video viewing context-aware mobile video quality adaptation method that leverages various IoT data sources, such as sensor data and image data. By analyzing this data and comprehending the intricate mapping between context and perceived quality, SenseQ is able to minimize network usage while preserving comparable quality perception effectively. In an experiment involving 20 participants, we evaluated the effect of context on HVS quality perception. We found that SenseQ has promising potential in achieving comparable perceptual quality while reducing network usage in IoT-enabled mobile video streaming systems.

Keywords

• Adaptive streaming
• neural networks
• machine learning
• mobile video QoE