Uniqueness filtering for local feature descriptors in urban building recognition

Giang Phuong Nguyen, Hans Jørgen Andersen

Publikation: Bidrag til tidsskriftKonferenceartikel i tidsskriftForskningpeer review

2 Citationer (Scopus)

Abstract

Existing local feature detectors such as Scale Invariant Feature Transform (SIFT) usually produce a large number of features per image. This is a major disadvantage in terms of the speed of search and recognition in a run-time application. Besides, not all detected features are equally important in the search. It is therefore essential to select informative descriptors. In this paper, we propose a new approach to selecting a subset of local feature descriptors. Uniqueness is used as a filtering criterion in selecting informative features. We formalize the notion of uniqueness and show how it can be used for selection purposes. To evaluate our approach, we carried out experiments in urban building recognition domains with different datasets. The results show a significant improvement not only in recognition speed, as a result of using fewer features, but also in the performance of the system with selected features.

OriginalsprogEngelsk
BogserieLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Vol/bind5099 LNCS
Sider (fra-til)85-93
Antal sider9
ISSN0302-9743
DOI
StatusUdgivet - 2008
Begivenhed3rd International Conference on Image and Signal Processing, ICISP 2008 - Cherbourg-Octeville, Frankrig
Varighed: 1 jul. 20083 jul. 2008

Konference

Konference3rd International Conference on Image and Signal Processing, ICISP 2008
Land/OmrådeFrankrig
ByCherbourg-Octeville
Periode01/07/200803/07/2008
SponsorInstitut Universitaire de Technologie Cherbourg Manche, Universite de Caen Basse-Normandie, ENSICAEN, Centre national de la recherche scientifique, Cent. Natl. de la Recherche Scientifique, Delegation Normandie

Fingeraftryk

Dyk ned i forskningsemnerne om 'Uniqueness filtering for local feature descriptors in urban building recognition'. Sammen danner de et unikt fingeraftryk.

Citationsformater