Kullback-Leibler Divergence based Marker Detection in Augmented Reality

A marker is a pre-defined pattern that is scanned for in augmented reality videos. Its presence when detected would provide the exact location for the placement of the 3D object in the video. The accurate detection of markers is based on efficient matching strategies, correlation being the most common measure. In this experimental study, we investigate the utility of the Kullback-Leibler (KL) divergence for marker matching and detection. This requires the marker intensities to be normalized to a probability distribution since the KL divergence measures the distance between two probability distributions.