I. Introduction
Deployments of large surveillance video networks are common in many application domains including law enforcement, critical infrastructure security, public utility monitoring and private institution security. Now, many environments such as shopping malls, city streets, educational campuses, banks, medical facilities, mass transports etc. are equipped with video cameras and supporting infrastructure. Though these camera networks are used for general monitoring purposes, they also sometimes capture video segments that hold significant information that can aid in investigations. In addition to video captured by fixed camera infrastructure, it is commonplace for bystanders to capture videos of incidents on their smartphones. Collectively the large amounts of video data captured from multiple sources have to be inspected when there is a need to investigate videos from those channels.