| At present, more and more multimedia applications are widely used, and multimedia technology is developing rapidly. Video coding technology exits everywhere in every aspect of daily life. With the development of high-definition technology, high-definition terminal display equipments are used more and more widely. But standard-definition video exits in all kinds of video resources. How to solve the problem that data transmitted is standard-definition video, but receiving equipments are for the high-definition video is a question now.The difference between standard-definition and high-definition video is frame rate and spatial resolution. This paper is a research on frame rate up-conversion in the process of converting from standard-definition video to high-definition video. Motion vector can be used in the process of frame rate up-conversion for the video which has been compressed. It may appear interpolation holes in the image, if the motion vector is directly used in frame rate up-conversion. Consequently, motion vector needs to be disposed. First, record the motion vectors got from decoding, and dispose motion vectors to get correct ones and then do the frame rate up-conversion. The interpolation is done between the known frames using interpolation formula, along motion vector. The quality of this decoding information used and compressed video oriented method is the same as the traditional motion compensation interpolation in both subjective and objective results. It is effective.In this paper, this decoding information used and compressed video oriented method and several frame rate up-conversion methods are combined separately. These methods include overlapped block motion compensation (OBMC), adaptive overlapped block motion compensation (AOBMC), and motion aligned auto-regression model (MAAR). These methods perform better than direct interpolation method which is depicted before. Because overlapped block interpolation is used for interpolating pixels on the block edge. It makes that not only the motion vector in current block but also the motion vectors in adjacent blocks are used in the interpolation. And the interpolation weights are precise in AR model. The experiment results show that the sensitive degree of motion vector is the main factor that affects the interpolation effect of this frame rate up-conversion method. |
PDF Full Text Request