提出一种快速有效的非冗余依赖图构建算法,记录所有网格顶点分裂操作之间的依赖关系,并删除其中的冗余依赖.在此基础上,提出了一种有效的全局分步等划分分组算法,通过初始划分和全局细化将依赖图划分为若干子图,最终将每个子图打包为一个分组.实验结果表明,该算法能够最小化分组之间的依赖性,从而有效地减少了分组丢失引起的解压与绘制延时,并形成良好的渐进绘制效果.
In this packetization scheme, we firstly construct a non-redundant directed acyclic graph (DAG) encoding the dependencies among the entire vertex split operations of the progressive mesh. Based on this scheme, a global graph equal partition algorithm is presented for partitioning the dependency graph with special local-circular features into several equal size sub-graphs, which is packed as packets. Briefly, the global partition algorithm consists of two main phases including initial partition phase and global refinement phase. Experimental results demonstrate that the proposed schemes can minimize the dependencies of the packets while transmitting the progressive mesh over unreliable network, thus decrease the total delay time of unpacking the packets and rendering at client caused by the packets losing, and maximize the rendering nodes so as to achieve better progressive rendering quality.