A Database Perspective on Large Scale High-Dimensional Indexing
Year: 2014
Type: preprint
Abstract: This manuscript gives an overview of several years ofresearch aiming at inventing and evaluating systems and techniques forbuilding efficient content-based retrieval systems managing extremelylarge databases of multimedia material. It has focused on themanagement of still images.Most of the work described in this manuscript has been achieved whilekeeping in mind a database perspective. The most important notion thatsheds a particular light on the high-dimensional indexing techniqueswe have been developing is the \emph{mandatory need for secondary storage}. In essence, secondary storage is needed to cope withfailures, which is a requirement for systems running in the realworld. Secondary storage is also needed to handle truly large scaledatasets.
Author: Laurent Amsaleg
Language: French
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2612843237