Video Stabilization through Dynamic Analysis of Frames Signatures

Digital stabilization is a key element in hand-held imaging devices in that it allows to enjoy the acquired sequences without disturbing jerkiness. In order to effectively remove the unwanted camera movements, it is necessary to estimate the global motion between adjacent frames. This paper presents a novel technique for estimating the global motion between frames using dynamic time warping techniques. The proposed algorithm assures robustness also in presence of illumination changes and moving objects.