An adaptive bilateral motion estimation algorithm and its hardware architecture

Abstract

In this paper, an adaptive bilateral motion estimation (ABIME) algorithm is proposed for frame rate up-conversion of high definition (HD) video. The proposed ABIME algorithm can be used as a refinement step after a true motion estimation algorithm. It refines the motion vector field between successive frames by employing a spiral search pattern and by adaptively assigning weights to candidate search locations. An adaptive early termination technique is also proposed for reducing the computational complexity of the ABIME algorithm. In addition, a high performance hardware architecture is proposed for implementing the ABIME algorithm. The proposed hardware uses an efficient memory organization and a novel data reuse scheme in order to reduce the memory bandwidth and control overhead. The proposed hardware is capable of processing 278 1920x1080 full HD frames per second (fps), therefore doubling the frame rate to 556 fps. The proposed early termination technique reduces the energy consumption of the proposed hardware by 29%.