Although block compressive sensing (BCS) makes it tractable to senselarge-sized images and video, its recovery performance has yet to besignificantly improved because its recovered images or video usually sufferfrom blurred edges, loss of details, and high-frequency oscillatory artifacts,especially at a low subrate. This paper addresses these problems by designing amodified total variation technique that employs multi-block gradientprocessing, a denoised Lagrangian multiplier, and patch-based sparserepresentation. In the case of video, the proposed recovery method is able toexploit both spatial and temporal similarities. Simulation results confirm theimproved performance of the proposed method for compressive sensing of imagesand video in terms of both objective and subjective qualities.
展开▼
