Community Photo Tagging: Engagement and Quality Study

Abstract

With today's dissemination of digital cameras, running events are usually well-presented in various photo sharing platforms. The number of photos from one medium-sized race can easily exceed a thousand or two, therefore, it is a tedious task for a participant to find photos where he/she is depicted. Bibtaggers is a crowd-based service allowing volunteers (usually, race participants) to collectively tag photos with race bib numbers, therefore enabling fast photo search. The paper presents results obtained using the service, specifically, the quality of tags collected and the user engagement data. These results are also compared to bib tags collected with a help of Amazon Mechanical Turk using similar procedure but different (monetary) incentive.