Lightweight Web3D Crowd Rendering for Online Massive Conferencing
Zhicheng Liu, Linzhou Li, Feng Tian, Jinyuan Jia · 2022 IEEE International Symposium on Mixed and Augmented Reality Adjunct (ISMAR-Adjunct) · 2022
Massive virtual online conferences have been encountering technical challenges and limitations, such as huge data volume, slow network transmission and low rendering frame rate. To tackle these limitations, in this paper we propose a Lightweight Online Virtual conferencE (LOVE) framework. A pseudo instantiation technique is first applied to reduce the weight of massive dynamic scenarios. LOVE realizes a progressive transmission of pseudo instanced objects, that speeds up the network loading speed of dynamic scenarios. Moreover, an intelligent UV mapping module for rendering pseudo instanced objects is developed to improve the diversity of dynamic crowd. Our experiments on several scenes of large crowd have demonstrated that LOVE is able to keep high FPS when the size of crowd increases.