【摘要】 提出一种几何模型分组渐进压缩算法·利用偏移细分曲面的模型表示方法,通过网格分块准则将模型的控制网格划分为多个chart·在对曲面的偏移量进行小波变换后,采用零树编码算法对每个chart上的小波系数压缩编码,生成多条相互独立的渐进压缩码流·该算法支持动态多分辨率解码,能够从压缩结果中直接获得模型的视点相关多分辨率表示·实验结果表明,该算法有利于提高多分辨率模型表示和绘制的效率,降低分布式系统的网络带宽需求·更多还原

【Abstract】 An algorithm for chart based progressive geometry compression is proposed. Geometric models are presented as displaced subdivision surfaces and their control meshes are partitioned into multiple charts. After the wavelet transformation of displacements, zerotree encoding algorithms are used to encode the wavelet coefficients of each chart separately, resulting in multiple progressively compressed streams. From the compression result, one can directly reconstruct geometric models with different resolutions at different parts as governed by view parameters. Experimental results show that the algorithm facilitates the multiresolution control and rendering of geometric models and can decrease the network bandwidth occupation of distributed rendering systems.更多还原