Graduate School and Research Center in Digital Sciences

Direct modeling of image keypoints distribution through copula-based image signatures

Redi, Miriam; Merialdo, Bernard

ICMR 2013, ACM International Conference on Multimedia Retrieval, April 16-19, Dallas, Texas, USA

Local Image Descriptors (LID) aggregation models such as Bag of Words and Fisher Vectors represent an image based on the distribution of its LIDs given a global model, e.g. a visual codebook or a Gaussian Mixture. Inspired by Copula theory, in this paper we propose a LID-based feature that represents directly the behavior of the image LID distribution, without requiring to compute a global model. Following the definition of Copula, we represent the distribution of the image LIDs by describing, on one side, its marginals, and on the other side, a Copula function. The Copula defines the dependencies between the marginals and their mapping to a multivariate probability distribution function. We test the resulting feature for scene recognition and video retrieval (Trecvid data), showing that our approach outperforms, in both tasks, the Bag of Words and the Fisher Vectors Model.

Document Doi Bibtex

Title:Direct modeling of image keypoints distribution through copula-based image signatures
Keywords:Scene Recognition, Feature Extraction, CBIR, Gaussian Copulae
Type:Conference
Language:English
City:Dallas
Country:UNITED STATES
Date:
Department:Data Science
Eurecom ref:3939
Copyright: © ACM, 2013. This is the author's version of the work. It is posted here by permission of ACM for your personal use. Not for redistribution. The definitive version was published in ICMR 2013, ACM International Conference on Multimedia Retrieval, April 16-19, Dallas, Texas, USA http://dx.doi.org/10.1145/2461466.2461498
Bibtex: @inproceedings{EURECOM+3939, doi = {http://dx.doi.org/10.1145/2461466.2461498}, year = {2013}, title = {{D}irect modeling of image keypoints distribution through copula-based image signatures}, author = {{R}edi, {M}iriam and {M}erialdo, {B}ernard}, booktitle = {{ICMR} 2013, {ACM} {I}nternational {C}onference on {M}ultimedia {R}etrieval, {A}pril 16-19, {D}allas, {T}exas, {USA}}, address = {{D}allas, {UNITED} {STATES}}, month = {04}, url = {http://www.eurecom.fr/publication/3939} }
See also: