Analysis of user behavior and difficulty in labeling polygons of a segmented image in a citizen science project

Abstract

Citizen science projects attempt to use knowledge, time or computing resources of volunteers to achieve a particular scientific objective. It fosters collaboration between scientists and the general public. Several citizen-science projects are already in execution, some of those successful. In particular, ’wisdom of the crowds’ may be used to get a consensual opinion on a particular topic or object (e.g. classification or labeling purposes), by collecting and analyzing opinions of users on those topics or objects. Citizen science projects collect data that is used for its stated purpose (e.g. object classification), but we consider that all the collected data may be used to give interesting insights on the reasoning of the project’s volunteers. In this paper we present a citizen-science project (volunteer labeling of imprecisely segmented image regions) and show which information can be inferred, through basic analysis, about individual and collective behavior of volunteers of the project.