An Effective Global Motion Estimation Method Based on a Compact Motion Model

This paper proposes a global motion estimation (GME) method based on a compact motion model to predict the changes between the successive frames in panoramic video coding. Compared with other complex motion models, the compact motion model uses only three motion parameters to describe the translational motion and non-translational motion (i.e. zooming) accurately in GME. Experimental results show that the proposed method is a low complexity algorithm, whose coding performance is up to 1.36dB compared with the conventional method for compressing the panoramic video.