| Title |
A Study on Edge Cloud-Based AR Split-Rendered Video Real-Time Streaming |
| Authors |
장준환(Junhwan Jang) ; 최민수(Minsu Choi) ; 이준석(Junsuk Lee) ; 윤사일(Sangpil Yoon) ; 양진욱(Jinwook Yang) ; 박우출(woochool Park) |
| DOI |
https://doi.org/10.5573/ieie.2021.58.6.71 |
| Keywords |
Split-rendering; Augmented reality; MPEG-DASH; Edge computing; Calibration |
| Abstract |
Previous AR(Augmented Reality)-based services have had issues with the expansion of range due to the necessity of expensive devices which have powerful processing power for spatial recognition, registration and rendering. In this paper, we implemented AR Split-Rendering system on edge server, which deals with all kinds of high computational works of AR contents instead of AR end-user devices. The AR Split-Rendering system was designed and implemented to provide AR services to as many users as possible. In addition, we brought pose prediction technology to reduce the registration error rate which is caused by network and processing delay by more than 90%, and finally confirmed that the real-time playback performance, 4K(3840x2160)@30fps and 2K(1920x1080)@60fps. |