Generating a synthetic video dataset

Abstract

A synthetic video dataset, scenario, and task were included in the 2009 VAST Challenge, to allow participants an opportunity to demonstrate visual analytic tool use on video data. This is the first time a video challenge had been presented as part of the VAST contest and provided interesting challenges in task and dataset development, video analytic tool development, and metrics for judging entries. We describe the considerations and requirements for generation of a usable challenge, the video creation itself, and some submissions and assessments from that mini-challenge.