Visual analogies: A framework for defining aspect categorization

P. Daphne Tsatsoulis, Bryan A. Plummer, David Alexander Forsyth

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Analogies are common simple word problems (calf is to cow as x is to sheep?) and we use them to identify analogies between images. Let I[A, θ] be an image of object A at view θ. We show how to learn to choose an image I such that I[A, φ] is to I[A, θ] as I is to I[B, θ]. We introduce a framework to identify an image of a familiar object at an unfamiliar angle and extend our method to treat unfamiliar objects. By doing so, we identify pairs of objects that are good at finding new views of one another. This yields an operational notion of aspectual equivalence: objects are equivalent if they can predict each other’s appearance well.

Original languageEnglish (US)
Title of host publicationComputer Vision – ECCV 2016 Workshops, Proceedings
EditorsGang Hua, Herve Jegou
PublisherSpringer-Verlag
Pages540-547
Number of pages8
ISBN (Print)9783319494081
DOIs
StatePublished - Jan 1 2016
Event14th European Conference on Computer Vision, ECCV 2016 - Amsterdam, Netherlands
Duration: Oct 11 2016Oct 14 2016

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9915 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other14th European Conference on Computer Vision, ECCV 2016
CountryNetherlands
CityAmsterdam
Period10/11/1610/14/16

Fingerprint

Categorization
Analogy
Word problem
Choose
Equivalence
Framework
Vision
Object
Angle
Predict

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Tsatsoulis, P. D., Plummer, B. A., & Forsyth, D. A. (2016). Visual analogies: A framework for defining aspect categorization. In G. Hua, & H. Jegou (Eds.), Computer Vision – ECCV 2016 Workshops, Proceedings (pp. 540-547). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9915 LNCS). Springer-Verlag. https://doi.org/10.1007/978-3-319-49409-8_47

Visual analogies : A framework for defining aspect categorization. / Tsatsoulis, P. Daphne; Plummer, Bryan A.; Forsyth, David Alexander.

Computer Vision – ECCV 2016 Workshops, Proceedings. ed. / Gang Hua; Herve Jegou. Springer-Verlag, 2016. p. 540-547 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9915 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Tsatsoulis, PD, Plummer, BA & Forsyth, DA 2016, Visual analogies: A framework for defining aspect categorization. in G Hua & H Jegou (eds), Computer Vision – ECCV 2016 Workshops, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9915 LNCS, Springer-Verlag, pp. 540-547, 14th European Conference on Computer Vision, ECCV 2016, Amsterdam, Netherlands, 10/11/16. https://doi.org/10.1007/978-3-319-49409-8_47
Tsatsoulis PD, Plummer BA, Forsyth DA. Visual analogies: A framework for defining aspect categorization. In Hua G, Jegou H, editors, Computer Vision – ECCV 2016 Workshops, Proceedings. Springer-Verlag. 2016. p. 540-547. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-49409-8_47
Tsatsoulis, P. Daphne ; Plummer, Bryan A. ; Forsyth, David Alexander. / Visual analogies : A framework for defining aspect categorization. Computer Vision – ECCV 2016 Workshops, Proceedings. editor / Gang Hua ; Herve Jegou. Springer-Verlag, 2016. pp. 540-547 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{18acfbbf87d549a5bc94bdcbee04e2a5,
title = "Visual analogies: A framework for defining aspect categorization",
abstract = "Analogies are common simple word problems (calf is to cow as x is to sheep?) and we use them to identify analogies between images. Let I[A, θ] be an image of object A at view θ. We show how to learn to choose an image I such that I[A, φ] is to I[A, θ] as I is to I[B, θ]. We introduce a framework to identify an image of a familiar object at an unfamiliar angle and extend our method to treat unfamiliar objects. By doing so, we identify pairs of objects that are good at finding new views of one another. This yields an operational notion of aspectual equivalence: objects are equivalent if they can predict each other’s appearance well.",
author = "Tsatsoulis, {P. Daphne} and Plummer, {Bryan A.} and Forsyth, {David Alexander}",
year = "2016",
month = "1",
day = "1",
doi = "10.1007/978-3-319-49409-8_47",
language = "English (US)",
isbn = "9783319494081",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer-Verlag",
pages = "540--547",
editor = "Gang Hua and Herve Jegou",
booktitle = "Computer Vision – ECCV 2016 Workshops, Proceedings",

}

TY - GEN

T1 - Visual analogies

T2 - A framework for defining aspect categorization

AU - Tsatsoulis, P. Daphne

AU - Plummer, Bryan A.

AU - Forsyth, David Alexander

PY - 2016/1/1

Y1 - 2016/1/1

N2 - Analogies are common simple word problems (calf is to cow as x is to sheep?) and we use them to identify analogies between images. Let I[A, θ] be an image of object A at view θ. We show how to learn to choose an image I such that I[A, φ] is to I[A, θ] as I is to I[B, θ]. We introduce a framework to identify an image of a familiar object at an unfamiliar angle and extend our method to treat unfamiliar objects. By doing so, we identify pairs of objects that are good at finding new views of one another. This yields an operational notion of aspectual equivalence: objects are equivalent if they can predict each other’s appearance well.

AB - Analogies are common simple word problems (calf is to cow as x is to sheep?) and we use them to identify analogies between images. Let I[A, θ] be an image of object A at view θ. We show how to learn to choose an image I such that I[A, φ] is to I[A, θ] as I is to I[B, θ]. We introduce a framework to identify an image of a familiar object at an unfamiliar angle and extend our method to treat unfamiliar objects. By doing so, we identify pairs of objects that are good at finding new views of one another. This yields an operational notion of aspectual equivalence: objects are equivalent if they can predict each other’s appearance well.

UR - http://www.scopus.com/inward/record.url?scp=85006067156&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85006067156&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-49409-8_47

DO - 10.1007/978-3-319-49409-8_47

M3 - Conference contribution

AN - SCOPUS:85006067156

SN - 9783319494081

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 540

EP - 547

BT - Computer Vision – ECCV 2016 Workshops, Proceedings

A2 - Hua, Gang

A2 - Jegou, Herve

PB - Springer-Verlag

ER -