Dataset Information

Calorie Estimation From Pictures of Food: Crowdsourcing Study.

ABSTRACT: BACKGROUND:Software designed to accurately estimate food calories from still images could help users and health professionals identify dietary patterns and food choices associated with health and health risks more effectively. However, calorie estimation from images is difficult, and no publicly available software can do so accurately while minimizing the burden associated with data collection and analysis. OBJECTIVE:The aim of this study was to determine the accuracy of crowdsourced annotations of calorie content in food images and to identify and quantify sources of bias and noise as a function of respondent characteristics and food qualities (eg, energy density). METHODS:We invited adult social media users to provide calorie estimates for 20 food images (for which ground truth calorie data were known) using a custom-built webpage that administers an online quiz. The images were selected to provide a range of food types and energy density. Participants optionally provided age range, gender, and their height and weight. In addition, 5 nutrition experts provided annotations for the same data to form a basis of comparison. We examined estimated accuracy on the basis of expertise, demographic data, and food qualities using linear mixed-effects models with participant and image index as random variables. We also analyzed the advantage of aggregating nonexpert estimates. RESULTS:A total of 2028 respondents agreed to participate in the study (males: 770/2028, 37.97%, mean body mass index: 27.5 kg/m2). Average accuracy was 5 out of 20 correct guesses, where "correct" was defined as a number within 20% of the ground truth. Even a small crowd of 10 individuals achieved an accuracy of 7, exceeding the average individual and expert annotator's accuracy of 5. Women were more accurate than men (P<.001), and younger people were more accurate than older people (P<.001). The calorie content of energy-dense foods was overestimated (P=.02). Participants performed worse when images contained reference objects, such as credit cards, for scale (P=.01). CONCLUSIONS:Our findings provide new information about how calories are estimated from food images, which can inform the design of related software and analyses.

SUBMITTER: Zhou J

PROVIDER: S-EPMC6246963 | biostudies-literature | 2018 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Calorie Estimation From Pictures of Food: Crowdsourcing Study.

Zhou Jun J Bell Dane D Nusrat Sabrina S Hingle Melanie M Surdeanu Mihai M Kobourov Stephen S

Interactive journal of medical research 20181105 2

<h4>Background</h4>Software designed to accurately estimate food calories from still images could help users and health professionals identify dietary patterns and food choices associated with health and health risks more effectively. However, calorie estimation from images is difficult, and no publicly available software can do so accurately while minimizing the burden associated with data collection and analysis.<h4>Objective</h4>The aim of this study was to determine the accuracy of crowdsour ...[more]

PMID: 30401671

Similar Datasets

Project description:Background and aimsIn addition to craving responses to salient food cues, the anticipation of short-term rewarding consumption of palatable food may overrun the anticipation of long-term negative consequences of obesity. The present investigation addressed the potential interplay of food cravings and decision-making abilities in individuals with obesity.MethodStudy 1 included 107 bariatric surgery candidates with class 2/3 obesity (OB-group) and study 2 included 54 individuals with normal weight/pre-obesity (nonOB-group). In both studies, standardized questionnaires concerning food cravings, food addiction, and psychopathology were administered. A cue-reactivity paradigm was used to measure craving responses toward semi-individualized images of highly palatable, processed food/fruit (appetitive food cues) compared to images of raw vegetables (non-appetitive food cues). Decision-making was measured with a modified computerized version of the Iowa Gambling Task (IGT) with food pictures. Both groups were divided into two subgroups that were randomized to different IGT conditions. In one IGT condition the advantageous IGT card decks were covered by pictures of palatable, processed food or fruit and the disadvantageous decks by images of raw vegetables (= congruent condition), and in the other IGT condition vice versa.ResultsParticipants in the OB-group admitted on average higher craving responses toward palatable, processed food or fruit cues compared to pictures of raw vegetables. This was not the case in the nonOB-group. Contrary to our hypothesis, decision-making performance in both groups was worse when pictures of palatable, processed food or fruit were associated with advantageous IGT card decks compared to performance when those pictures were linked to the disadvantageous decks. The interference effect of food pictures processing on advantageous decision-making has been observed particularly in those individuals of the OB-group who exhibited high craving responses toward palatable, processed food cues or high levels of food addiction.DiscussionThe results indicate that food pictures processing interferes with decision-making, regardless of weight status. Opposed to the hypothesis, stronger tendencies to avoid than to approach pictures presenting processed, tasty food were observed. Further research should examine how cognitive avoidance tendencies toward processed, high energy food and approach tendencies toward healthy food can be transferred to real life situations.

Dataset Information

Calorie Estimation From Pictures of Food: Crowdsourcing Study.

Publications

Calorie Estimation From Pictures of Food: Crowdsourcing Study.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets