On affine invariant descriptors related to SIFT

Using a classical result on algebraic invariants of the unimodular group, we present in this paper some basic geometric affine invariant quantities, and we use them to construct some distinctive descriptors for object detection. Although full affine invariance cannot be guaranteed due to noncommutat...

Descripción completa

Detalles Bibliográficos
Autores: Sadek, Rida, Constantinopoulos, Constantinos, Meinhardt Llopis, Enric, Ballester, Coloma, Caselles, Vicente
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2012
País:España
Institución:Universitat Pompeu Fabra
Repositorio:Repositorio Digital de la UPF
OAI Identifier:oai:repositori.upf.edu:10230/46739
Acceso en línea:http://hdl.handle.net/10230/46739
http://dx.doi.org/10.1137/100798739
Access Level:acceso abierto
Palabra clave:Image matching
Affine invariance
Image descriptors
Object recognition
Descripción
Sumario:Using a classical result on algebraic invariants of the unimodular group, we present in this paper some basic geometric affine invariant quantities, and we use them to construct some distinctive descriptors for object detection. Although full affine invariance cannot be guaranteed due to noncommutativity of camera blur with affine maps and the domain problem (that is, the difficulty of finding an affine covariant domain), the proposed descriptors behave more robustly than SIFT with respect to affine deformations. This is supported by our comparisons both with the version of SIFT computed on an affine normalized neighborhood, and with ASIFT, which solves both the previously mentioned camera blur and domain problems by cleverly sampling the orbit of affine transformations of the images.