PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Adapted vocabularies for generic visual categorization
Florent Perronnin, Chris Dance, Gabriela Csurka and Marco Bressan
In: Computer Vision – ECCV 2006 Lecture Notes in Computer Science (3954). (2006) Springer Berlin / Heidelberg , pp. 464-475. ISBN 978-3-540-33838-3

Abstract

Several state-of-the-art Generic Visual Categorization (GVC) systems are built around a vocabulary of visual terms and characterize images with one histogram of visual word counts. We propose a novel and practical approach to GVC based on a universal vocabulary, which describes the content of all the considered classes of images, and class vocabularies obtained through the adaptation of the universal vocabulary using class-specific data. An image is characterized by a set of histograms - one per class - where each histogram describes whether the image content is best modeled by the universal vocabulary or the corresponding class vocabulary. It is shown experimentally on three very different databases that this novel representation outperforms those approaches which characterize an image with a single histogram.

EPrint Type:Book Section
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Machine Vision
ID Code:2251
Deposited By:Gabriela Csurka
Deposited On:11 October 2006