A Block-Grouping Method for Image Denoising by Block Matching and 3-D Transform Filtering

Document Type: Research Paper


University of Sistan and Baluchestan


Image denoising by block matching and threedimensional
transform filtering (BM3D) is a two steps state-ofthe-
art algorithm that uses the redundancy of similar blocks in
noisy image for removing noise. Similar blocks which can have
some overlap are found by a block matching method and grouped
to make 3-D blocks for 3-D transform filtering. In this paper we
propose a new block grouping algorithm in the first step of
BM3D that improves the performance of denoising algorithm
especially in heavy noise conditions. In heavy noise conditions,
BM3D causes some artifacts in the filtered image. These artifacts
are reduced by the proposed block grouping algorithm. In the
proposed block grouping method, beside of a similarity measure
used for block matching, the amount of overlap between blocks is
considered. Experimental results show that the proposed block
grouping method can improve the performance of BM3D in
terms of both peak signal-to-noise ratio (PSNR) and visual