Data-dependent metric filtering
作者：
Highlights：
• This article is focused on the similarity searching in generic metric spaces.
• We consider embedding of pairwise distances to make triangles in 2D Euclidean space.
• We analyse angles in triangles to improve bounds given by triangle inequalities.
• The analysis enables to treat pivots independently to achieve top quality filtering.
• We strongly improve the metric filtering with a negligible loss in effectiveness.
摘要
•This article is focused on the similarity searching in generic metric spaces.•We consider embedding of pairwise distances to make triangles in 2D Euclidean space.•We analyse angles in triangles to improve bounds given by triangle inequalities.•The analysis enables to treat pivots independently to achieve top quality filtering.•We strongly improve the metric filtering with a negligible loss in effectiveness.
论文关键词：68P10,65D05,65D17,Metric Space Searching,Similarity Search,Metric Filtering,Data Dependent Filtering
论文评审过程：Received 17 February 2021, Revised 13 August 2021, Accepted 20 December 2021, Available online 23 December 2021, Version of Record 12 May 2022.
论文官网地址：https://doi.org/10.1016/j.is.2021.101980
