Abstract
Body measurements, including weight and height, are key indicators of health. Being able to visually assess body measurements reliably is a step towards increased awareness of overweight and obesity and is thus important for public health. Nevertheless it is currently not well understood how accurately humans can assess weight and height from images, and when and how they fail. To bridge this gap, we start from 1,682 images of persons collected from the Web, each annotated with the true weight and height, and ask crowd workers to estimate the weight and height for each image. We conduct a faceted analysis taking into account characteristics of the images as well as the crowd workers assessing the images, revealing several novel findings: (1) Even after aggregation, the crowd’s accuracy is overall low. (2) We find strong evidence of contraction bias toward a reference value, such that the weight of light people and the height of short people are overestimated, whereas the weight of heavy people and the height of tall people are underestimated. (3) We estimate workers’ individual reference values using a Bayesian model, finding that reference values strongly correlate with workers’ own height and weight, indicating that workers are better at estimating people similar to themselves. (4) The weight of tall people is underestimated more than that of short people; yet, knowing the height decreases the weight error only mildly. (5) Accuracy is higher on images of females than of males, but female and male workers are no different in terms of accuracy. (6) Crowd workers improve over time if given feedback on previous guesses. Finally, we explore various bias correction models for improving the crowd’s accuracy, but find that this only leads to modest gains. Overall, this work provides important insights on biases in body measurement estimation as obesity-related conditions are on the rise.
Highlights
Height and weight play a key role as indicators of health [1]
We found that various post-hoc bias correction models only led to modest gains, whereas varying the setup of the crowdsourcing task itself is more promising: the accuracy could be improved by giving workers feedback on the quality of their guesses, effectively teaching them to make better guesses in the future
We develop a simple generative model that describes the process of guessing and is based on the notions of reference values and contraction bias introduced in Sect
Summary
Height and weight play a key role as indicators of health [1]. Population-level height and weight statistics open a window onto the general health of a region. We observed that the guesses collected for both images follow similar distributions, and even come from similar pools of workers (in terms of the workers’ own weight and height distributions) Under such conditions, no statistical model can correct the results. Given that height does not change significantly during adulthood, whereas weight may change widely and might have to be frequently re-estimated for the same person, one might provide workers with the true height labels for the images and collect only weight guesses For this setup, we apply an additional filter based on workers’ countries of residence: as previously shown, reference values of height for people in Asian countries do not match the dataset at hand. The previous image, appears to be important when workers start to receive feedback: in almost all cases, the average accuracy of the guesses improved when the new image was similar to the previous one
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
