Efficient Tile-Based Adaptive Streaming for 360-Degree Video-on-Demand Systems

In the field of 360-degree video streaming, the optimization of visual quality within the viewer's field of view, alongside the minimization of system resource consumption, is a critical concern. The considerable file sizes associated with high-quality 360-degree videos pose significant challenges for efficient transmission. Current strategies primarily employ a tile-based approach to address these issues. However, these approaches face two core challenges: the methodology for tile segmentation, and the strategy for downloading tiles. To address these challenges, this paper presents a tile-based solution designed specifically for adaptive 360-degree video-on-demand streaming. Our method surpasses baseline methods in extensive simulations, positioning it as one of the leading solutions in the MMSys'24 Grand Challenge1.