Multimedia Systems Computer Networks and Communications
CCUCB-FR: An optimization approach for edge-client collaborative foveated rendering in virtual reality
Abstract
The rendering of immersive virtual reality (VR) frames currently demands significant computational power to provide users with an optimal VR experience. Edge computing technology is anticipated to offer computational support for VR rendering. However, the latency associated with the rendering and transmission of high-quality VR frames continues to disrupt immersion and the sense of presence in interactive VR experiences. To address this challenge, this paper leverages the visual characteristics of the human eye, dynamically dividing the rendering task into two regions: the foveal region rendered on the user’s device and the peripheral vision region rendered at the edge. A collaborative rendering framework based on foveated pre-rendering is proposed to minimize end-to-end latency while maintaining the quality of experience. Furthermore, this paper presents an optimization problem that considers network bandwidth constraints, rendering frame complexity, and device rendering capabilities within the edge-client collaborative rendering environment. This problem is modeled as a Combinatorial Multi-Armed Bandit problem, and the Contextual Combinatorial Upper Confidence Bound for Foveated Rendering algorithm is introduced to solve it. Finally, the experimental results demonstrate the effectiveness of the proposed algorithm.