用代数格实现点数据索引

Indexing Point Data by Lattice

提出用代数格实现多维点数据索引的一般方法 .在该方法中 ,首先用格矢量量化器对点数据进行量化 ,然后用倒排文件和散列表存储和索引量化后的点数据 ,以便利用代数格良好的代数和几何性质实现快速检索 .最后给出了基于 E8格的八维点数据索引性能的计算机模拟结果 .

A general method to index multi dimensional point data is first proposed in this paper. In this method, lattice based vector quantizer is first adopted to quantize high dimensional point data, the inverted file and Hash table are then used to store and index the quantized point data, and fast retrieval is implemented at last because of the good geometric and algebraic properties of the lattice. In the experiment, lattice E 8 is used as an example to index large scale 8 dimension point data, and the indexing performance is obtained by computer simulation.