Part 1:  Introduction
Part 2:  Video Basics

The preliminary version of the roadmap was presented at the SPIE Europe Security Defence 2008 conference but was not published in the proceedings.

The roadmap presented here represents the collective input from researchers and practitioners worldwide.  It represents a consensus of the ideas put forth based upon each contributor’s video analysis specialization.  In an attempt to make the roadmap manageable the research topics have been grouped under three heading:

• Content extraction – develop capabilities to analyze video content and produce an understanding of events taking place
• Intelligent content services – develop capabilities to manage and provide automated access to video content
• Enabling technologies – develop capabilities that apply in both content extraction and services and build rudimentary video multimodal fusion to support complex reasoning on video content

Please note that none of the headings or research topics is cast in stone and is subject to change based on feedback I receive.

Before delving into the roadmap I would like to offer an explanation of what you will find and how to interpret it.  Let’s call each research topic a swim lane.  As you move to the right in the swim lane expect that the solution matures (improved precision and recall) and then tackles harder and harder problems associated with the swim lane technology.  Blue labeling in a swim lane indicates it was a technology identified in an earlier version of the roadmap.  Red labeled swim lanes indicate a new technology topic that was added.  Red labeling in swim lanes indicates an additional level of analytical processing capability.  The timeline dates are estimates based on the perceived difficulty of research achieving a solution and, more importantly, assume that the level of resource support to the necessary research is available, which is basically a bad assumption whose impact will only move the completion milestone to the right.