Feature map hashing: Sub-linear indexing of appearance and global geometry

Avrithis, Y; Tolias, G; Kalantidis, Y

dc.contributor.author	Avrithis, Y	en
dc.contributor.author	Tolias, G	en
dc.contributor.author	Kalantidis, Y	en
dc.date.accessioned	2014-03-01T02:46:47Z
dc.date.available	2014-03-01T02:46:47Z
dc.date.issued	2010	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/32856
dc.subject	feature maps	en
dc.subject	hashing	en
dc.subject	image retrieval	en
dc.subject	indexing geometry	en
dc.subject	sub-linear indexing	en
dc.subject.other	Background clutter	en
dc.subject.other	Excellent performance	en
dc.subject.other	Feature map	en
dc.subject.other	Global geometries	en
dc.subject.other	hashing	en
dc.subject.other	Image alignment	en
dc.subject.other	Image geometries	en
dc.subject.other	Image indexing and retrieval	en
dc.subject.other	indexing geometry	en
dc.subject.other	Indexing process	en
dc.subject.other	Inverted files	en
dc.subject.other	Local feature	en
dc.subject.other	Min-wise independent permutations	en
dc.subject.other	New approaches	en
dc.subject.other	Photometric variations	en
dc.subject.other	Query time	en
dc.subject.other	Retrieval process	en
dc.subject.other	Set intersection	en
dc.subject.other	Shape parameters	en
dc.subject.other	Similarity measure	en
dc.subject.other	sub-linear indexing	en
dc.subject.other	Visual word	en
dc.subject.other	Computational geometry	en
dc.subject.other	Image retrieval	en
dc.subject.other	Indexing (of information)	en
dc.title	Feature map hashing: Sub-linear indexing of appearance and global geometry	en
heal.type	conferenceItem	en
heal.identifier.primary	10.1145/1873951.1873985	en
heal.identifier.secondary	http://dx.doi.org/10.1145/1873951.1873985	en
heal.publicationDate	2010	en
heal.abstract	We present a new approach to image indexing and retrieval, which integrates appearance with global image geometry in the indexing process, while enjoying robustness against viewpoint change, photometric variations, occlusion, and background clutter. We exploit shape parameters of local features to estimate image alignment via a single correspondence. Then, for each feature, we construct a sparse spatial map of all remaining features, encoding their normalized position and appearance, typically vector quantized to visual word. An image is represented by a collection of such feature maps and RANSAC-like matching is reduced to a number of set intersections. Because the induced dissimilarity is still not a metric, we extend min-wise independent permutations to collections of sets and derive a similarity measure for feature map collections. We then exploit sparseness to build an inverted file whereby the retrieval process is sub-linear in the total number of images, ideally linear in the number of relevant ones. We achieve excellent performance on 104 images, with a query time in the order of milliseconds. © 2010 ACM.	en
heal.journalName	MM'10 - Proceedings of the ACM Multimedia 2010 International Conference	en
dc.identifier.doi	10.1145/1873951.1873985	en
dc.identifier.spage	231	en
dc.identifier.epage	240	en