Image upsampling via spatially adaptive block-matching filtering
Aram Danielyan, Alessandro Foi, Vladimir Y. Katkovnik, Karen Egiazarian · 2008
In this paper we present a novel wavelet-domain image upsam-pling algorithm based on iterative spatially adaptive Þltering. A high-resolution image is reconstructed by alternating two proce-dures: spatially adaptive Þltering and projection on the observation-constrained subspace. The Block-Matching and 3D Þltering (BM3D) [3] technique is used to suppress ringing, and reconstruct missing wavelet detail coefÞcients. The BM3D algorithm exploits the local image statistics collected from similar blocks to extract local and non-local image features by 3D transform-domain shrink-age. It results in high-quality upsampled images, with sharp edges and practically no artifacts. 1.