An efficient representation for multi-application accessible visual information

Abstract

As computers become more and more pervasive in our society, one of the problems that arises is how to efficiently integrate different types of systems and the information that they collect. Through cooperation between multi-modal systems, a much broader range of applications become available. In an attempt to attack this problem on a small scale, this project demonstrates the coupling of a natural language processing system to a camera-based person tracker for the purpose of being able to ask informally worded questions about what the tracker has observed in the MIT Intelligent Room.