Video Coding With Key Frames Guided Super-Resolution

In this paper a video coding scheme with Layered Block Matching Super-resolution (LBM-SR) is presented. At the encoder side, it divides the video frames into key and non-key frames, which are encoded at original resolution and reduced resolution respectively. During the resolution reduction process, most of the high frequency information in non-key frames is dropped to save the bit-rate. At the decoder side, LBM-SR utilizes a Layered Block Matching method in wavelet domain to restore the lost high frequency parts of the non-key frame, with the nearby key frames as a reference. Due to the similarity between key frames and non-key frames, the experimental result is remarkable and the whole scheme is demonstrated to be a promising one.