An Interactive Light Field Video System with User-Dependent View Selection and Coding Scheme

摘要

Due to the sheer size of Light Field (LF) video, the traffic of LF video is many times larger than traditional multimedia, which brings new challenges on how to efficiently store and transmit this huge amount of complex data. In this paper, we propose an interactive LF video streaming system with user-dependent view selection and coding scheme. The user trajectory prediction model proposed in this paper can be used to calculate the viewing area of the user in a limited consecutive number of time slots. With the projection model, the system can determine the interested views of the user for transmission. Furthermore, we construct a special LF video sequence with only the selected view at a consecutive number of time slots. By reducing the number of views for transmission, the system can significantly reduce the traffic load in transmission and realize a low-latency real-time LF video application.