Towards bottom-up analysis of social food

Jaclyn Rich, Hamed Haddadi, Timothy M. Hospedales

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

4 Citations (Scopus)

Abstract

Social media provide a wealth of information for research into public health by providing a rich mix of personal data, location, hashtags, and social network information. Among these, Instagram has been recently the subject of many computational social science studies. However despite Instagram's focus on image sharing, most studies have exclusively focused on the hashtag and social network structure. In this paper we perform the first large scale content analysis of Instagram posts, addressing both the image and the associated hashtags, aiming to understand the content of partiallylabelled images taken in-The-wild and the relationship with hashtags that individuals use as noisy labels. In particular, we explore the possibility of learning to recognise food image content in a data driven way, discovering both the categories of food, and how to recognise them, purely from social network data. Notably, we demonstrate that our approach to food recognition can often achieve accuracies greater than 70% in recognising popular food-related image categories, despite using no manual annotation. We highlight the current capabilities and future challenges and opportunities for such data-driven analysis of image content and the relation to hashtags.

Original languageEnglish
Title of host publicationDH 2016 - Proceedings of the 2016 Digital Health Conference
PublisherAssociation for Computing Machinery, Inc
Pages111-120
Number of pages10
ISBN (Print)9781450342247
DOIs
Publication statusPublished - 11 Apr 2016
Externally publishedYes
Event6th International Conference on Digital Health, DH 2016 - Montreal, Canada
Duration: 11 Apr 201613 Apr 2016

Other

Other6th International Conference on Digital Health, DH 2016
CountryCanada
CityMontreal
Period11/4/1613/4/16

Fingerprint

Food Analysis
Social Support
Food
Social Media
Data privacy
Social sciences
Social Sciences
Public health
Labels
Public Health
Learning
Research

Keywords

  • Image Recognition
  • Instagram
  • Machine Learning
  • Social Media

ASJC Scopus subject areas

  • Health Information Management
  • Computer Science Applications
  • Health Informatics

Cite this

Rich, J., Haddadi, H., & Hospedales, T. M. (2016). Towards bottom-up analysis of social food. In DH 2016 - Proceedings of the 2016 Digital Health Conference (pp. 111-120). Association for Computing Machinery, Inc. https://doi.org/10.1145/2896338.2897734

Towards bottom-up analysis of social food. / Rich, Jaclyn; Haddadi, Hamed; Hospedales, Timothy M.

DH 2016 - Proceedings of the 2016 Digital Health Conference. Association for Computing Machinery, Inc, 2016. p. 111-120.

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

Rich, J, Haddadi, H & Hospedales, TM 2016, Towards bottom-up analysis of social food. in DH 2016 - Proceedings of the 2016 Digital Health Conference. Association for Computing Machinery, Inc, pp. 111-120, 6th International Conference on Digital Health, DH 2016, Montreal, Canada, 11/4/16. https://doi.org/10.1145/2896338.2897734
Rich J, Haddadi H, Hospedales TM. Towards bottom-up analysis of social food. In DH 2016 - Proceedings of the 2016 Digital Health Conference. Association for Computing Machinery, Inc. 2016. p. 111-120 https://doi.org/10.1145/2896338.2897734
Rich, Jaclyn ; Haddadi, Hamed ; Hospedales, Timothy M. / Towards bottom-up analysis of social food. DH 2016 - Proceedings of the 2016 Digital Health Conference. Association for Computing Machinery, Inc, 2016. pp. 111-120
@inproceedings{66f1938f48f449a2ab9878e78da2c5a2,
title = "Towards bottom-up analysis of social food",
abstract = "Social media provide a wealth of information for research into public health by providing a rich mix of personal data, location, hashtags, and social network information. Among these, Instagram has been recently the subject of many computational social science studies. However despite Instagram's focus on image sharing, most studies have exclusively focused on the hashtag and social network structure. In this paper we perform the first large scale content analysis of Instagram posts, addressing both the image and the associated hashtags, aiming to understand the content of partiallylabelled images taken in-The-wild and the relationship with hashtags that individuals use as noisy labels. In particular, we explore the possibility of learning to recognise food image content in a data driven way, discovering both the categories of food, and how to recognise them, purely from social network data. Notably, we demonstrate that our approach to food recognition can often achieve accuracies greater than 70{\%} in recognising popular food-related image categories, despite using no manual annotation. We highlight the current capabilities and future challenges and opportunities for such data-driven analysis of image content and the relation to hashtags.",
keywords = "Image Recognition, Instagram, Machine Learning, Social Media",
author = "Jaclyn Rich and Hamed Haddadi and Hospedales, {Timothy M.}",
year = "2016",
month = "4",
day = "11",
doi = "10.1145/2896338.2897734",
language = "English",
isbn = "9781450342247",
pages = "111--120",
booktitle = "DH 2016 - Proceedings of the 2016 Digital Health Conference",
publisher = "Association for Computing Machinery, Inc",

}

TY - GEN

T1 - Towards bottom-up analysis of social food

AU - Rich, Jaclyn

AU - Haddadi, Hamed

AU - Hospedales, Timothy M.

PY - 2016/4/11

Y1 - 2016/4/11

N2 - Social media provide a wealth of information for research into public health by providing a rich mix of personal data, location, hashtags, and social network information. Among these, Instagram has been recently the subject of many computational social science studies. However despite Instagram's focus on image sharing, most studies have exclusively focused on the hashtag and social network structure. In this paper we perform the first large scale content analysis of Instagram posts, addressing both the image and the associated hashtags, aiming to understand the content of partiallylabelled images taken in-The-wild and the relationship with hashtags that individuals use as noisy labels. In particular, we explore the possibility of learning to recognise food image content in a data driven way, discovering both the categories of food, and how to recognise them, purely from social network data. Notably, we demonstrate that our approach to food recognition can often achieve accuracies greater than 70% in recognising popular food-related image categories, despite using no manual annotation. We highlight the current capabilities and future challenges and opportunities for such data-driven analysis of image content and the relation to hashtags.

AB - Social media provide a wealth of information for research into public health by providing a rich mix of personal data, location, hashtags, and social network information. Among these, Instagram has been recently the subject of many computational social science studies. However despite Instagram's focus on image sharing, most studies have exclusively focused on the hashtag and social network structure. In this paper we perform the first large scale content analysis of Instagram posts, addressing both the image and the associated hashtags, aiming to understand the content of partiallylabelled images taken in-The-wild and the relationship with hashtags that individuals use as noisy labels. In particular, we explore the possibility of learning to recognise food image content in a data driven way, discovering both the categories of food, and how to recognise them, purely from social network data. Notably, we demonstrate that our approach to food recognition can often achieve accuracies greater than 70% in recognising popular food-related image categories, despite using no manual annotation. We highlight the current capabilities and future challenges and opportunities for such data-driven analysis of image content and the relation to hashtags.

KW - Image Recognition

KW - Instagram

KW - Machine Learning

KW - Social Media

UR - http://www.scopus.com/inward/record.url?scp=84966692515&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84966692515&partnerID=8YFLogxK

U2 - 10.1145/2896338.2897734

DO - 10.1145/2896338.2897734

M3 - Conference contribution

SN - 9781450342247

SP - 111

EP - 120

BT - DH 2016 - Proceedings of the 2016 Digital Health Conference

PB - Association for Computing Machinery, Inc

ER -