Real-Time High-Quality Visualization for Volumetric Contents Rendering: A Lyapunov Optimization Framework

Abstract

Real-time volumetric contents streaming on augmented reality (AR) devices should necessitate a balance between end-users' quality of experience (QoE) and the latency requirements. Lowering the quality of the volumetric contents to diminish the latency hinders the user's QoE. Otherwise, setting the quality of volumetric contents relatively high to improve the users' QoE increases the latency, which can be challenging to meet user satisfaction in AR services. Based on this trade-off observation, our proposed method maximizes time-average AR quality under latency requirements, inspired by Lyapunov optimization framework. In order to control the AR quality depending on latency requirements, we control the point cloud rendering ratio in the volumetric contents under the concept of Lyapunov optimization. Our extensive evaluation demonstrates that our proposed method achieves desired performance improvements, i.e., avoiding latency growing while ensuring the high quality of the volumetric contents streaming in AR services.