With today’s dynamic multimedia collections, maintenance of high-dimensional indexes is an important, yet understudied topic. Extended Cluster Pruning (eCP) is a highly-scalable approximate indexing approach based on clustering, that is targeted at stable performance in a disk-based scenario. In this work, we propose an index maintenance strategy for the eCP index, which utilizes the tree structure of the index and its approximate nature. We then develop a cost model for the strategy and evaluate its cost using a simulation model.
|Titel||Proceedings of the International Conference on Similarity Search and Applications (SISAP)|
|Status||Udgivet - okt. 2019|
|Navn||Lecture Notes in Computer Science|