基于参数化混合元球体表示的高分辨率DTI纤维丛可视化

Visualizing High-Resolution DTI Fibers with Parametric Merging Metaballs

为了克服传统DTI可视化技术存在计算空间大、信息丢失等问题,提出一种基于Perfect Spatial Hashing稀疏数据压缩方法的高分辨率DTI纤维丛可视化方法.将流场可视化中的streamball表示改进为沿着积分曲线布局的可参数化混合元球体,并将这种参数化混合元球体表示规范为一个高分辨率的稀疏三维密度场;进而采用Perfect Spatial Hashing稀疏数据压缩方法压缩该密度场,在保持数据高精度的同时提供了数据的高效随机访问特性.实验结果表明,采用文中方法得到的可视化结果不仅能清晰地揭示组织结构的连通性,还能展示局部张量细节信息.用户只需简单地改变等值面参数就可实时观察可视化结果.

We present a novel method for visualizing high-resolution DTI fibers with merging metaballs to cope with the problem of conventional DTI visualization approaches such as memory consuming,information losing.We extend the streamball in flow visualization to tensor merging metaballs and place them along integral curves,yielding a sparse 3D density influence field.To represent this sparse influence field,we use perfect spatial hashing method to compress it while retaining efficient data access.The resulting visualization shows that our approach can not only reveal the connectivity information in biological tissue,but also the local tensor details.By interactively changing the iso-value parameter,users can easily exploit the real-time visualization results.