Abstract
Among researches of video super-resolution (VSR), reconstruction-based method is widely used for the capability of managing arbitrary scenes. And for accuracy, full-image based motion prediction is often adopted that results outstanding outputs. But it costs huge time complexity due to the multi-level processing. In this paper, we propose an edge-based motion and intensity prediction scheme to reduce the computation cost while maintain good enough quality simultaneously. The key point of reducing computation cost is to focus on extracted edges rather than the whole frame when finding optical flows of the video sequence in accordance with human vision system (HVS). Bi-directional optical flows are usually adopted to increase the prediction accuracy but it also increase the computation time. We also propose to obtain the backward flow from foregoing forward flow prediction which effectively save the heavy load. We perform a series of experiments and comparisons between existing VSR methods and our proposed edge-based method with different sequences and upscaling factors. The results reveal that our proposed scheme can successfully keep the super-resolved sequence quality and get about five times speed up in computation time for a two by two video scaling and six times speed up for a four by four one.