Research On Orthogonal Range Query Based On BC-iDistance

With the advent of the era of big data,people are paying more attention to the content and value of data.For multimedia information such as images,audio,and video,the traditional search methods can no longer meet people's needs.kNN queries,range queries,and orthogonal range queries are common content-based query methods.For efficient querying,data is usually stored in an index that is optimized for either kNN or range query.This can be problematic when data is optimized for kNN retrieval and a user needs a range query or vice versa.The paper focuses on the advanced BC-iDistance index.By analyzing the problems of the na?ve method and the spatial decomposition method,a new method based on data decomposition is proposed.Without changing the original data structure and indexing mechanism of BC-iDistance,the BC-iDistance index used for kNN queries and range queries can support orthogonal range queries.The main research content of this paper is as follows:(1)The na?ve method introduces a large number of unsuitable vectors in candidates which leading to a terrible query efficiency,the paper proposes a method based on data decomposition.The method focuses on the area where the orthogonal range query area intersects with the cluster,and ignores those parts of the query area that have no data in the orthogonal range.The orthogonal range query was implemented without changing the original data structure and indexing mechanism of BC-iDistance.The experiment uses different distributed data sets to perform orthogonal range queries under multiple query conditions,demonstrates the practicability of data decomposition methods,and proves the feasibility of constructing orthogonal range queries on kNN or range query indexes.(2)The orthogonal range query algorithm W-Chord in P2 P network is proposed for the performance bottleneck problem caused by the standalone version of the index.W-Chord combines orthogonal range query algorithm based on data decomposition with Chord protocol to realize orthogonal range query under P2 P network.(3)Based on the above techniques,an image similarity search subsystem was developed.The subsystem provides both range query and orthogonal range query to achieve image similarity retrieval.