JCISE Application Paper Crowdsourcing Canonical Views of 3D Models

This paper reports the initial results of employing a commercial Crowdsourcing (aka Micro-outsourcing) service to carryout a geometric reasoning task (i.e. identification of a characteristic view) that is typical of the ill defined problems found in many CAD/CAM applications. Although Crowdsourcing sites (which distribute browser based tasks to potentially large numbers of anonymous workers on the Internet) are well established for image analysis and text manipulation there is little academic work on the effectiveness or limitations of the approach when applied to 2 and 3D geometry. The work reported here describes the initial results of using Crowdsourcing to determine the “best ” canonical, or characteristic, views of complex 3D models of engineering components. The results suggest that the approach is a cheap, fast and effective method of providing pragmatic solutions to difficult computational problems. This paper is concerned with the combination of two approaches to problem solving