Learning to localize little landmarks

Saurabh Singh, Derek Hoiem, David Forsyth

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

Abstract

We interact everyday with tiny objects such as the door handle of a car or the light switch in a room. These little landmarks are barely visible and hard to localize in images. We describe a method to find such landmarks by finding a sequence of latent landmarks, each with a prediction model. Each latent landmark predicts the next in sequence, and the last localizes the target landmark. For example, to find the door handle of a car, our method learns to start with a latent landmark near the wheel, as it is globally distinctive, subsequent latent landmarks use the context from the earlier ones to get closer to the target. Our method is supervised solely by the location of the little landmark and displays strong performance on more difficult variants of established tasks and on two new tasks.

Original languageEnglish (US)
Title of host publicationProceedings - 29th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016
PublisherIEEE Computer Society
Pages260-269
Number of pages10
ISBN (Electronic)9781467388504
DOIs
StatePublished - Dec 9 2016
Event29th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016 - Las Vegas, United States
Duration: Jun 26 2016Jul 1 2016

Publication series

NameProceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition
Volume2016-December
ISSN (Print)1063-6919

Conference

Conference29th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016
CountryUnited States
CityLas Vegas
Period6/26/167/1/16

Fingerprint

Door handles
Railroad cars
Wheels
Switches

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition

Cite this

Singh, S., Hoiem, D., & Forsyth, D. (2016). Learning to localize little landmarks. In Proceedings - 29th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016 (pp. 260-269). [7780404] (Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition; Vol. 2016-December). IEEE Computer Society. https://doi.org/10.1109/CVPR.2016.35

Learning to localize little landmarks. / Singh, Saurabh; Hoiem, Derek; Forsyth, David.

Proceedings - 29th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016. IEEE Computer Society, 2016. p. 260-269 7780404 (Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition; Vol. 2016-December).

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

Singh, S, Hoiem, D & Forsyth, D 2016, Learning to localize little landmarks. in Proceedings - 29th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016., 7780404, Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol. 2016-December, IEEE Computer Society, pp. 260-269, 29th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016, Las Vegas, United States, 6/26/16. https://doi.org/10.1109/CVPR.2016.35
Singh S, Hoiem D, Forsyth D. Learning to localize little landmarks. In Proceedings - 29th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016. IEEE Computer Society. 2016. p. 260-269. 7780404. (Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition). https://doi.org/10.1109/CVPR.2016.35
Singh, Saurabh ; Hoiem, Derek ; Forsyth, David. / Learning to localize little landmarks. Proceedings - 29th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016. IEEE Computer Society, 2016. pp. 260-269 (Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition).
@inproceedings{ee62e38f2d294d5baab318036cde687f,
title = "Learning to localize little landmarks",
abstract = "We interact everyday with tiny objects such as the door handle of a car or the light switch in a room. These little landmarks are barely visible and hard to localize in images. We describe a method to find such landmarks by finding a sequence of latent landmarks, each with a prediction model. Each latent landmark predicts the next in sequence, and the last localizes the target landmark. For example, to find the door handle of a car, our method learns to start with a latent landmark near the wheel, as it is globally distinctive, subsequent latent landmarks use the context from the earlier ones to get closer to the target. Our method is supervised solely by the location of the little landmark and displays strong performance on more difficult variants of established tasks and on two new tasks.",
author = "Saurabh Singh and Derek Hoiem and David Forsyth",
year = "2016",
month = "12",
day = "9",
doi = "10.1109/CVPR.2016.35",
language = "English (US)",
series = "Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition",
publisher = "IEEE Computer Society",
pages = "260--269",
booktitle = "Proceedings - 29th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016",

}

TY - GEN

T1 - Learning to localize little landmarks

AU - Singh, Saurabh

AU - Hoiem, Derek

AU - Forsyth, David

PY - 2016/12/9

Y1 - 2016/12/9

N2 - We interact everyday with tiny objects such as the door handle of a car or the light switch in a room. These little landmarks are barely visible and hard to localize in images. We describe a method to find such landmarks by finding a sequence of latent landmarks, each with a prediction model. Each latent landmark predicts the next in sequence, and the last localizes the target landmark. For example, to find the door handle of a car, our method learns to start with a latent landmark near the wheel, as it is globally distinctive, subsequent latent landmarks use the context from the earlier ones to get closer to the target. Our method is supervised solely by the location of the little landmark and displays strong performance on more difficult variants of established tasks and on two new tasks.

AB - We interact everyday with tiny objects such as the door handle of a car or the light switch in a room. These little landmarks are barely visible and hard to localize in images. We describe a method to find such landmarks by finding a sequence of latent landmarks, each with a prediction model. Each latent landmark predicts the next in sequence, and the last localizes the target landmark. For example, to find the door handle of a car, our method learns to start with a latent landmark near the wheel, as it is globally distinctive, subsequent latent landmarks use the context from the earlier ones to get closer to the target. Our method is supervised solely by the location of the little landmark and displays strong performance on more difficult variants of established tasks and on two new tasks.

UR - http://www.scopus.com/inward/record.url?scp=84986259962&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84986259962&partnerID=8YFLogxK

U2 - 10.1109/CVPR.2016.35

DO - 10.1109/CVPR.2016.35

M3 - Conference contribution

AN - SCOPUS:84986259962

T3 - Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition

SP - 260

EP - 269

BT - Proceedings - 29th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016

PB - IEEE Computer Society

ER -