Object class recognition by unsupervised scale-invariant learning

DMCA

Object class recognition by unsupervised scale-invariant learning (2003)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by R. Fergus , P. Perona , A. Zisserman

Venue:	In CVPR
Citations:	1125 - 50 self

BibTeX

@INPROCEEDINGS{Fergus03objectclass,
    author = {R. Fergus and P. Perona and A. Zisserman},
    title = {Object class recognition by unsupervised scale-invariant learning},
    booktitle = {In CVPR},
    year = {2003},
    pages = {264--271}
}

OpenURL

Abstract

We present a method to learn and recognize object class models from unlabeled and unsegmented cluttered scenes in a scale invariant manner. Objects are modeled as flexible constellations of parts. A probabilistic representation is used for all aspects of the object: shape, appearance, occlusion and relative scale. An entropy-based feature detector is used to select regions and their scale within the image. In learning the parameters of the scale-invariant object model are estimated. This is done using expectation-maximization in a maximum-likelihood setting. In recognition, this model is used in a Bayesian manner to classify images. The flexible nature of the model is demonstrated by excellent results over a range of datasets including geometrically constrained classes (e.g. faces, cars) and flexible objects (such as animals). 1.

Keyphrases

unsupervised scale-invariant learning object class recognition relative scale maximum-likelihood setting object class model flexible nature entropy-based feature detector scale invariant manner bayesian manner flexible object flexible constellation unsegmented cluttered scene scale-invariant object model probabilistic representation excellent result

DMCA