Skip to Main content Skip to Navigation
Conference papers

Object localization by subspace clustering of local descriptors

Charles Bouveyron 1 Juho Kannala 2 Cordelia Schmid 1, * Stéphane Girard 3 
* Corresponding author
1 LEAR - Learning and recognition in vision
GRAVIR - IMAG - Laboratoire d'informatique GRAphique, VIsion et Robotique de Grenoble, Inria Grenoble - Rhône-Alpes, CNRS - Centre National de la Recherche Scientifique : FR71
Abstract : This paper presents a probabilistic approach for object localization which combines subspace clustering with the selection of discriminative clusters. Clustering is often a key step in object recognition and is penalized by the high dimensionality of the descriptors. Indeed, local descriptors, such as SIFT, which have shown excellent results in recognition, are high-dimensional and live in different low-dimensional subspaces. We therefore use a subspace clustering method called High-Dimensional Data Clustering (HDDC) which overcomes the curse of dimensionality. Furthermore, in many cases only a few of the clusters are useful to discriminate the object. We, thus, evaluate the discriminative capacity of clusters and use it to compute the probability that a local descriptor belongs to the object. Experimental results demonstrate the effectiveness of our probabilistic approach for object localization and show that subspace clustering gives better results compared to standard clustering methods. Furthermore, our approach outperforms existing results for the Pascal 2005 dataset.
Document type :
Conference papers
Complete list of metadata
Contributor : THOTH Team Connect in order to contact the contributor
Submitted on : Monday, December 20, 2010 - 9:49:38 AM
Last modification on : Thursday, January 20, 2022 - 5:30:13 PM

Links full text




Charles Bouveyron, Juho Kannala, Cordelia Schmid, Stéphane Girard. Object localization by subspace clustering of local descriptors. 5th Indian Conference on Computer Vision, Graphics and Image Processing (ICVGIP '06), Dec 2006, Madurai, India. pp.457-467, ⟨10.1007/11949619_41⟩. ⟨inria-00548589⟩



Record views