Learning object categories from internet image searches

Rob Fergus, Li Fei-Fei, Pietro Perona, Andrew Zisserman

Research output: Contribution to journalArticle

Abstract

In this paper, we describe a simple approach to learning models of visual object categories from images gathered from Internet image search engines. The images for a given keyword are typically highly variable, with a large fraction being unrelated to the query term, and thus pose a challenging environment from which to learn. By training our models directly from Internet images, we remove the need to laboriously compile training data sets, required by most other recognition approachesthis opens up the possibility of learning object category models on-the-fly. We describe two simple approaches, derived from the probabilistic latent semantic analysis (pLSA) technique for text document analysis, that can be used to automatically learn object models from these data. We show two applications of the learned model: first, to rerank the images returned by the search engine, thus improving the quality of the search engine; and second, to recognize objects in other image data sets.

Original languageEnglish (US)
Article number5483225
Pages (from-to)1453-1466
Number of pages14
JournalProceedings of the IEEE
Volume98
Issue number8
DOIs
StatePublished - Aug 2010

Fingerprint

Internet
Search engines
Semantics

Keywords

  • Internet image search engines
  • Learning
  • Object categories
  • Recognition
  • Unsupervised

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Learning object categories from internet image searches. / Fergus, Rob; Fei-Fei, Li; Perona, Pietro; Zisserman, Andrew.

In: Proceedings of the IEEE, Vol. 98, No. 8, 5483225, 08.2010, p. 1453-1466.

Research output: Contribution to journalArticle

Fergus, R, Fei-Fei, L, Perona, P & Zisserman, A 2010, 'Learning object categories from internet image searches', Proceedings of the IEEE, vol. 98, no. 8, 5483225, pp. 1453-1466. https://doi.org/10.1109/JPROC.2010.2048990
Fergus, Rob ; Fei-Fei, Li ; Perona, Pietro ; Zisserman, Andrew. / Learning object categories from internet image searches. In: Proceedings of the IEEE. 2010 ; Vol. 98, No. 8. pp. 1453-1466.
@article{998f7a492dde4ed5b093dc9f4c02dd24,
title = "Learning object categories from internet image searches",
abstract = "In this paper, we describe a simple approach to learning models of visual object categories from images gathered from Internet image search engines. The images for a given keyword are typically highly variable, with a large fraction being unrelated to the query term, and thus pose a challenging environment from which to learn. By training our models directly from Internet images, we remove the need to laboriously compile training data sets, required by most other recognition approachesthis opens up the possibility of learning object category models on-the-fly. We describe two simple approaches, derived from the probabilistic latent semantic analysis (pLSA) technique for text document analysis, that can be used to automatically learn object models from these data. We show two applications of the learned model: first, to rerank the images returned by the search engine, thus improving the quality of the search engine; and second, to recognize objects in other image data sets.",
keywords = "Internet image search engines, Learning, Object categories, Recognition, Unsupervised",
author = "Rob Fergus and Li Fei-Fei and Pietro Perona and Andrew Zisserman",
year = "2010",
month = "8",
doi = "10.1109/JPROC.2010.2048990",
language = "English (US)",
volume = "98",
pages = "1453--1466",
journal = "Proceedings of the IEEE",
issn = "0018-9219",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "8",

}

TY - JOUR

T1 - Learning object categories from internet image searches

AU - Fergus, Rob

AU - Fei-Fei, Li

AU - Perona, Pietro

AU - Zisserman, Andrew

PY - 2010/8

Y1 - 2010/8

N2 - In this paper, we describe a simple approach to learning models of visual object categories from images gathered from Internet image search engines. The images for a given keyword are typically highly variable, with a large fraction being unrelated to the query term, and thus pose a challenging environment from which to learn. By training our models directly from Internet images, we remove the need to laboriously compile training data sets, required by most other recognition approachesthis opens up the possibility of learning object category models on-the-fly. We describe two simple approaches, derived from the probabilistic latent semantic analysis (pLSA) technique for text document analysis, that can be used to automatically learn object models from these data. We show two applications of the learned model: first, to rerank the images returned by the search engine, thus improving the quality of the search engine; and second, to recognize objects in other image data sets.

AB - In this paper, we describe a simple approach to learning models of visual object categories from images gathered from Internet image search engines. The images for a given keyword are typically highly variable, with a large fraction being unrelated to the query term, and thus pose a challenging environment from which to learn. By training our models directly from Internet images, we remove the need to laboriously compile training data sets, required by most other recognition approachesthis opens up the possibility of learning object category models on-the-fly. We describe two simple approaches, derived from the probabilistic latent semantic analysis (pLSA) technique for text document analysis, that can be used to automatically learn object models from these data. We show two applications of the learned model: first, to rerank the images returned by the search engine, thus improving the quality of the search engine; and second, to recognize objects in other image data sets.

KW - Internet image search engines

KW - Learning

KW - Object categories

KW - Recognition

KW - Unsupervised

UR - http://www.scopus.com/inward/record.url?scp=77954860650&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77954860650&partnerID=8YFLogxK

U2 - 10.1109/JPROC.2010.2048990

DO - 10.1109/JPROC.2010.2048990

M3 - Article

AN - SCOPUS:77954860650

VL - 98

SP - 1453

EP - 1466

JO - Proceedings of the IEEE

JF - Proceedings of the IEEE

SN - 0018-9219

IS - 8

M1 - 5483225

ER -