Video Registration Based on SIFT Feature Vectors

Video registration is a difficult task especially when spurious frame intensity differences and spatial variations between the two frames are present. To robust Video registration algorithms to such spurious variations it can be useful to employ a video registration matching criteria on higher dimensional feature spaces. This paper will present an overview of our recent work on Video registration using high dimensional Video features and Scale Invariant Feature Vector (SIFT) Feature Vector matching criteria. New approach estimates of information divergence measures will be presented. We will demonstrate the advantage of our approach for Video registration.