Design and analysis of a video motion predictor

Abstract

21st century has experienced a vast range of video application, from internet streaming to HDTV broadcasting. However, video data consumes a large amount of storage capacity and require high bandwidth to transfer wirelessly. Thus, these data need to be compressed. Motion estimation plays a vital role in compressing these video data efficiently to achieve low power and real-time application. This paper will discuss the design and analysis of motion estimation architectures implemented using UMHexagonS fast search algorithm. The results show that the architectures can perform motion prediction efficiently to achieve real-time performance.