Finger Tracking as an Input Device for Augmented Reality

Abstract

This paper describes experiments in the use of cross-correlation as a means of tracking pointing devices for a digital desk. The first section introduces the motivation by describing the potential for applying real time computer vision to man machine interaction. The problem of tracking is then formulated and addressed as a problem of optimal signal detection. Signal detection leads to a formulation of tracking using cross-correlation with a reference template as shown section 2. The problems of normalisation, choosing the size of the reference template and search region are addressed. A method is provided to detect when to initiate tracking as well as when tracking has failed. The problem of updating the reference mask is also addressed. 1 Computer Vision and Man Machine Interaction. One of the effects of the continued exponential growth in available computing power has been an exponential decrease in the cost of hardware for real time computer vision. This trend has been accelerated...