Study of virtual octree model for volume data
面向体数据的虚拟八叉树模型研究

Pointer-based octree and linear octree have their limitations either on time efficiency or storage efficiency. A pointerless/non-location-code virtual octree model based on multi-z-order-curves was proposed. In terms of time, the virtual octree provides fast access to the octree nodes by regularly partitioned node blocks and their index. In terms of storage, the virtual octree provides good compression by pointerless/non-location-code nodes and the node combination method based on regularly partitioned blocks. Test results show that the virtual octree has access efficiency as fast as pointer-based octree, and storage compression as good as linear octree. It is a good volume data model, and has wide application in volume graphics related areas.