Fast searching in image databases using multi-index robust fingerprinting

Abstract

Robust image fingerprinting seeks to transform a given input image into a compact binary hash using a non-invertible transform. These binary hashes exhibit robustness against common image processing and find their extensive application in multimedia databases where near neighbor index search is often employed. Unfortunately, robust fingerprinting length is usually longer than 32 bits which makes impossible to use them as direct indices in multimedia databases. This paper analyses a theoretical approach that allows to map a rneighbor search in Hamming space into a couple of direct index searches, using multiple hash tables built on fingerprinting substrings. We analyse the performances of this approach using a concrete perceptual fingerprinting scheme that we previously detailed in other paper. Experimental results conducted on a well known 4000 image dataset confirm dramatic speed-ups over a linear scan approach.