Understanding Image-based Big Data using Human Computation

This project will enhance machine understanding of big image and video data sets by including human input as sub-routines inside of larger computational systems. There are many existing marketplaces for micro-work, where people are able to bid and complete small-scale jobs, and humans are much better than machines at resolving issues such as whether a circular object is a face or a wheel. However, building computational systems that contain human feedback is not trivial. This research will focus on characterizing the results of specific human computation platforms and on developing instructions and algorithms that will best incorporate human computation in ways that allow robust computer understanding of large image-based datasets, such as in cancer detection and treatment.