An improved adaptive rood pattern search algorithm based on temporal and spatial correlation
Chunchun Chen, Junming Shan, Eryan Yang · 2012
Motion estimation (ME) is the most complicated and time-consuming part in the process of frame rate up-conversion (FRUC). In order to improve the speed and accuracy of ME, especially the speed and accuracy of motion vector (MV) prediction, an improved Adaptive Rood Pattern Search (ARPS) algorithm is proposed in this paper. Temporal and spatial correlation of frames and blocks is used in this paper. Based on ARPS-2 algorithm, it enhances the prediction of starting search point and decision of motion type. It also utilizes spatial correlation of frame to reduce the computational time. Compared with ARPS-2 algorithm, it improves the search efficiency, while ensures search accuracy.