An Extended Small Diamond Search Algorithm for Fast Block Motion Estimation

In this paper, we propose a fast motion estimation algorithm that is suitable for searching both center-biased motion and large motion by applying the small diamond pattern used in some block-matching algorithms (BMA). The search will be terminated by a small diamond search (SDS) method after performing a large search based on a modified three-step search (3SS) strategy. The results of the experiment show an increase about 220% in the search speed compared to that of diamond search (DS) and efficient three-step search (E3SS) in a sequence that represent small motions in objects. Experimental results also demonstrate reasonable search points in the estimation of rough motions in high-resolution images, maintaining a performance better than other fast BMAs in terms of PSNR.