Designing human computation games to update street views and its possibility

Abstract

Although the Web has abundant information, it does not necessarily contain the latest, most recentlyupdated information. In particular, interactive map websites and the accompanying street viewapplications often have outdated information because street views change constantly and are very costlyto update. In this work, we propose the LOX (Labeling and O/X) framework { a scalable human computationframework that discovers the latest updates of street views through mobile games. The challengein this work is providing an interface to identify the differences between the latest street view and theoutdated street view images. The LOX framework addresses this problem through three mobile games{ a compatibility ESP (cESP) Game, a Drag&Drop Labeling Game and an O/X Game. The cESP andLabeling Games encourage users to identify new updates on outdated street view images while the O/XGame increases the precision of the differences identified by the users between the existing street viewimages and the latest street view. We conducted several user studies to assess the design and usability ofthe game. The collected data allowed us to assess the performance of utilizing human computation gamesto update new information while providing us guidance on how to design user interfaces and workowsto increase the quality of the data. The games provide an opportunity for significant cost savings tothe service providers by providing an inexpensive method to determine which street views need to beupdated.