High dimensional spatial data query method, device and computer readable medium

Abstract

The invention provides a high dimensional spatial data query method, a device and a computer readable medium. The high dimensional spatial data query method comprises the following steps: determining a corresponding Z value for each data point of a high dimensional space; sequentially storing data points to disc pages in a disk according to the size order of corresponding Z values; establishing an index table according to the corresponding Z value of each data point and each disk page storing the data points; determining a query point and calculating the distance between each disk page in the index table and the query point; and carrying out data query according to the distance between each disk page in the index table and the query point. The query efficiency can be improved by the scheme.

Description

technical field [0001] The invention relates to the field of query technology, in particular to a query method, device and computer-readable medium for high-dimensional spatial data. Background technique [0002] With the explosive growth of multimedia data, the retrieval of massive information has encountered unprecedented challenges. Since multimedia data is mostly unstructured data, including images, videos, and audio, its semantic information cannot be directly expressed by underlying features, making traditional information retrieval no longer applicable, and it is urgent to develop retrieval methods for multimedia data. [0003] The usual practice is to extract feature vectors for multimedia data elements, and transform similarity retrieval on massive multimedia databases into a NN (Nearest Neighbor, nearest neighbor) query problem among large-scale high-dimensional feature vectors. Given a query object, use the specified metric to measure the similarity between featu...

AI Technical Summary

Problems solved by technology

[0004] However, when performing similarity query in the prior art, the query is random

Images