Abstract
BM3D is a state-of-the-art image denoising method. Its denoised results in the regions with strong edges can often be better than in the regions with smooth or weak edges, due to more accurate block-matching for the strong-edge regions. So using adaptive block sizes on different image regions may result in better image denoising. Based on these observations, in this paper, we first partition each image into regions belonging to one of the three morphological components, i.e., contour, texture, and smooth components, according to the regional energy of alternating current (AC) coefficients of discrete cosine transform (DCT). Then, we can adaptively determine the block size for each morphological component. Specifically, we use the smallest block size for the contour components, the medium block size for the texture components, and the largest block size for the smooth components. To better preserve image details, we also use a multi-stage strategy to implement image denoising, where every stage is similar to the BM3D method, except using adaptive sizes and different transform dimensions. Experimental results show that our proposed algorithm can achieve higher PSNR and MSSIM values than the BM3D method, and also better visual quality of denoised images than by the BM3D method and some other existing state-of-the-art methods.
Highlights
Image denoising, as a basic topic of pattern recognition and computer vision, has been studied for many years
We can see that both Peak signal to noise ratio (PSNR) and Mean structural similarity (MSSIM) values of SA-BM1-3D are consistently higher than those existing state-of-the-art algorithms, and are mostly higher than those of the block-matching 3D (BM3D)-SAPCA and weighted nuclear norm minimization (WNNM) algorithms
We can see that when the noise level is relatively high, SA-BM1-3D still hardly introduces any artifacts, whereas the BM3D algorithm introduces a lot of periodic artifacts
Summary
As a basic topic of pattern recognition and computer vision, has been studied for many years. BM3D is the current state-of-the-art image denoising method [3, 4], and, as a Recently, a variety of new image denoising methods have been proposed, but very few approaches can perform better than BM3D Many of these methods are based on the non-local idea, i.e., using similar image blocks (or patches) to explore new image denoising methods. Zhang et al [5] proposed a two-stage principal component analysis (PCA) on local pixel grouping, with its local pixel grouping achieved by block matching This method successfully combined the classical PCA with nonlocal idea and achieved better results than those traditional local methods, the denoising performance of this method is still lower than the BM3D method. Rajwade et al [6] used the higher order singular value
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
