| Summary: | Large-scale collaborative virtual environment allows multiple users connected to each other through the internet. Due to large amount of data needed to construct the complete scene, it’s very slow for a client host to replicate to all those data locally and render the whole scene. Although some existing systems had been developed for on-demand transmission and dynamic scene construction, there are still many technical issues and needs further improvements.
This project is directed at improving the pre-caching mechanism. In this dissertation, and approach of dynamic pre-caching mechanism based on visual importance factor is proposed. The mechanism partitions the virtual space into many zones and cells. And for each cell, the visual importance factor had been pre-calculated for all the relevant objects in the scene. Based on client user’s current state, those objects with high visibility importance factor are downloaded first and ready to be rendered into the scene. Also certain additional visible objects from the neighboring cell will be re-cached and ready to be rendered. This gives the client user a better virtual tour by reducing the latency that user needs to wait for visibility calculation, downloading and rendering object.
|
|---|