Scalable Remote Rendering using Synthesized Image Quality Assessment

Abstract

Depth-image-based rendering (DIBR) is widely used to support 3D interactive graphics on low-end mobile devices. Although it reduces the rendering cost on a mobile device, it essentially turns such a cost into depth image transmission cost or bandwidth consumption, inducing performance bottleneck to a remote rendering system. To address this problem, we design a scalable remote rendering framework based on synthesized image quality assessment. Specially, we design an efficient synthesized image quality metric based on Just Noticeable Distortion (JND), properly measuring human perceived geometric distortions in synthesized images. Based on this, we predict quality-aware reference viewpoints, with viewpoint intervals optimized by the JND-based metric. An adaptive transmission scheme is also developed to control depth image transmission based on perceived quality and network bandwidth availability. Experiment results show that our approach effectively reduces transmission frequency and network bandwidth consumption with perceived quality on mobile devices maintained. A prototype system is implemented to demonstrate the scalability of our proposed framework to multiple clients.