Estimating heights from photo collections

A data-driven approach

Ratan Dey, Madhurya Nangia, Keith Ross, Yong Liu

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

Abstract

A photo can potentially reveal a tremendous amount of information about an individual, including the individual's height, weight, gender, ethnicity, hair color, skin condition, interests, and wealth. A photo collection - a set of inter-related photos including photos of many people appearing in two or more photos - could potentially reveal a more vivid picture of the individuals in the collection. In this paper we consider the problem of estimating the heights of all the users in a photo collection, such as a collection of photos from a social network. The main ideas in our methodology are (i) for each individual photo, estimate the height differences among the people standing in the photo, (ii) from the photo collection, create a people graph, and combine this graph with the height difference estimates from the individual photos to generate height difference estimates among all the people in the collection, (iii) then use these height difference estimates, as well as an a priori distribution, to estimate the heights of all the people in the photo collection. Because many people will appear in multiple photos across the collection, height-difference estimates can be chained together, potentially reducing the errors in the estimates. To this end, we formulate a Maximum Likelihood Estimation (MLE) problem, which we show can be easily solved as a quadratic programming problem. Intuitively, this data-driven approach will improve as the number of photos and people in the collection increases. We apply the technique to estimating the heights of over 400 movie stars in the IMDb database and of about 30 graduate students.

Original languageEnglish (US)
Title of host publicationCOSN 2014 - Proceedings of the 2014 ACM Conference on Online Social Networks
PublisherAssociation for Computing Machinery, Inc
Pages227-238
Number of pages12
ISBN (Print)9781450331982
DOIs
StatePublished - Oct 1 2014
Event2nd ACM Conference on Online Social Networks, COSN 2014 - Dublin, Ireland
Duration: Oct 1 2014Oct 2 2014

Other

Other2nd ACM Conference on Online Social Networks, COSN 2014
CountryIreland
CityDublin
Period10/1/1410/2/14

Fingerprint

Maximum likelihood estimation
Quadratic programming
Stars
Skin
Students
Color

Keywords

  • Concept extraction
  • Height estimate
  • Image processing
  • Maximum likelihood estimation
  • People graph
  • Photo collection
  • Privacy

ASJC Scopus subject areas

  • Computer Networks and Communications

Cite this

Dey, R., Nangia, M., Ross, K., & Liu, Y. (2014). Estimating heights from photo collections: A data-driven approach. In COSN 2014 - Proceedings of the 2014 ACM Conference on Online Social Networks (pp. 227-238). Association for Computing Machinery, Inc. https://doi.org/10.1145/2660460.2660466

Estimating heights from photo collections : A data-driven approach. / Dey, Ratan; Nangia, Madhurya; Ross, Keith; Liu, Yong.

COSN 2014 - Proceedings of the 2014 ACM Conference on Online Social Networks. Association for Computing Machinery, Inc, 2014. p. 227-238.

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

Dey, R, Nangia, M, Ross, K & Liu, Y 2014, Estimating heights from photo collections: A data-driven approach. in COSN 2014 - Proceedings of the 2014 ACM Conference on Online Social Networks. Association for Computing Machinery, Inc, pp. 227-238, 2nd ACM Conference on Online Social Networks, COSN 2014, Dublin, Ireland, 10/1/14. https://doi.org/10.1145/2660460.2660466
Dey R, Nangia M, Ross K, Liu Y. Estimating heights from photo collections: A data-driven approach. In COSN 2014 - Proceedings of the 2014 ACM Conference on Online Social Networks. Association for Computing Machinery, Inc. 2014. p. 227-238 https://doi.org/10.1145/2660460.2660466
Dey, Ratan ; Nangia, Madhurya ; Ross, Keith ; Liu, Yong. / Estimating heights from photo collections : A data-driven approach. COSN 2014 - Proceedings of the 2014 ACM Conference on Online Social Networks. Association for Computing Machinery, Inc, 2014. pp. 227-238
@inproceedings{1648394dee594858aaa790f1ac5edd72,
title = "Estimating heights from photo collections: A data-driven approach",
abstract = "A photo can potentially reveal a tremendous amount of information about an individual, including the individual's height, weight, gender, ethnicity, hair color, skin condition, interests, and wealth. A photo collection - a set of inter-related photos including photos of many people appearing in two or more photos - could potentially reveal a more vivid picture of the individuals in the collection. In this paper we consider the problem of estimating the heights of all the users in a photo collection, such as a collection of photos from a social network. The main ideas in our methodology are (i) for each individual photo, estimate the height differences among the people standing in the photo, (ii) from the photo collection, create a people graph, and combine this graph with the height difference estimates from the individual photos to generate height difference estimates among all the people in the collection, (iii) then use these height difference estimates, as well as an a priori distribution, to estimate the heights of all the people in the photo collection. Because many people will appear in multiple photos across the collection, height-difference estimates can be chained together, potentially reducing the errors in the estimates. To this end, we formulate a Maximum Likelihood Estimation (MLE) problem, which we show can be easily solved as a quadratic programming problem. Intuitively, this data-driven approach will improve as the number of photos and people in the collection increases. We apply the technique to estimating the heights of over 400 movie stars in the IMDb database and of about 30 graduate students.",
keywords = "Concept extraction, Height estimate, Image processing, Maximum likelihood estimation, People graph, Photo collection, Privacy",
author = "Ratan Dey and Madhurya Nangia and Keith Ross and Yong Liu",
year = "2014",
month = "10",
day = "1",
doi = "10.1145/2660460.2660466",
language = "English (US)",
isbn = "9781450331982",
pages = "227--238",
booktitle = "COSN 2014 - Proceedings of the 2014 ACM Conference on Online Social Networks",
publisher = "Association for Computing Machinery, Inc",

}

TY - GEN

T1 - Estimating heights from photo collections

T2 - A data-driven approach

AU - Dey, Ratan

AU - Nangia, Madhurya

AU - Ross, Keith

AU - Liu, Yong

PY - 2014/10/1

Y1 - 2014/10/1

N2 - A photo can potentially reveal a tremendous amount of information about an individual, including the individual's height, weight, gender, ethnicity, hair color, skin condition, interests, and wealth. A photo collection - a set of inter-related photos including photos of many people appearing in two or more photos - could potentially reveal a more vivid picture of the individuals in the collection. In this paper we consider the problem of estimating the heights of all the users in a photo collection, such as a collection of photos from a social network. The main ideas in our methodology are (i) for each individual photo, estimate the height differences among the people standing in the photo, (ii) from the photo collection, create a people graph, and combine this graph with the height difference estimates from the individual photos to generate height difference estimates among all the people in the collection, (iii) then use these height difference estimates, as well as an a priori distribution, to estimate the heights of all the people in the photo collection. Because many people will appear in multiple photos across the collection, height-difference estimates can be chained together, potentially reducing the errors in the estimates. To this end, we formulate a Maximum Likelihood Estimation (MLE) problem, which we show can be easily solved as a quadratic programming problem. Intuitively, this data-driven approach will improve as the number of photos and people in the collection increases. We apply the technique to estimating the heights of over 400 movie stars in the IMDb database and of about 30 graduate students.

AB - A photo can potentially reveal a tremendous amount of information about an individual, including the individual's height, weight, gender, ethnicity, hair color, skin condition, interests, and wealth. A photo collection - a set of inter-related photos including photos of many people appearing in two or more photos - could potentially reveal a more vivid picture of the individuals in the collection. In this paper we consider the problem of estimating the heights of all the users in a photo collection, such as a collection of photos from a social network. The main ideas in our methodology are (i) for each individual photo, estimate the height differences among the people standing in the photo, (ii) from the photo collection, create a people graph, and combine this graph with the height difference estimates from the individual photos to generate height difference estimates among all the people in the collection, (iii) then use these height difference estimates, as well as an a priori distribution, to estimate the heights of all the people in the photo collection. Because many people will appear in multiple photos across the collection, height-difference estimates can be chained together, potentially reducing the errors in the estimates. To this end, we formulate a Maximum Likelihood Estimation (MLE) problem, which we show can be easily solved as a quadratic programming problem. Intuitively, this data-driven approach will improve as the number of photos and people in the collection increases. We apply the technique to estimating the heights of over 400 movie stars in the IMDb database and of about 30 graduate students.

KW - Concept extraction

KW - Height estimate

KW - Image processing

KW - Maximum likelihood estimation

KW - People graph

KW - Photo collection

KW - Privacy

UR - http://www.scopus.com/inward/record.url?scp=84912098034&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84912098034&partnerID=8YFLogxK

U2 - 10.1145/2660460.2660466

DO - 10.1145/2660460.2660466

M3 - Conference contribution

SN - 9781450331982

SP - 227

EP - 238

BT - COSN 2014 - Proceedings of the 2014 ACM Conference on Online Social Networks

PB - Association for Computing Machinery, Inc

ER -