Region-of-Interest Real-Time Tracking and Projection System

Abstract

This paper addresses the idea and implementation aspects of a real-time system for projection of image/video content onto moving object (human) regions of interest (ROIs) using a projector, based on object motion detection and tracking using fixed optical camera in an indoor environment. The developed system captures live video, extracts different body parts i.e. regions of interest in real time and makes an appropriate projection. We have developed an agent-based software architecture, that is intended to be easily extended and modified, where autonomous modules are cascaded to form a pipeline. Dynamic control mechanisms to enhance performance are also described. Furthermore, the use of a single, low cost optical camera, without the need for a complex sensor equipment can significantly decrease the overall system cost. In addition, users have more interaction with video contents and get more immersive experience and enjoyment. Preliminary experiments have shown that vivid results can be generated and evaluators are enganged with the framework presented herein. To the best of our knowledge, there have been no studies or work reported for the above mentioned application to be realized.