Image denoising by low‐rank approximation with estimation of noise energy distribution in SVD domain

Adaptive iterative global image denoising method based on SVD

“…Set12 contains two different sizes of images: and . In order to verify the effectiveness of our method, the new method is compared with six advanced methods such as BM3D method [17], GSR_NLS method [32], WNNM method [25], ASC_PCA method [22], LAR_SVD method [26] and NED_SVD method [28]. For a more intuitive comparison with six methods, we divide them into two categories and compare them separately.…”

“…SVD technology as an important matrix decomposition in linear algebra is introduced into image [39] and many related papers have appeared [25, 26, 28] (Fig. 1).…”

“…Fan et al . [28] uses the specific functional relationship between the original image and the noise image to remove noise. The papers [29, 30] proposed a denoising method for hyperspectral images.…”

Image Denoising Using Hybrid Singular Value Thresholding Operators

“…Since the noise statistics of the three channels are different, a weight matrix is introduced to balance their data fidelity. Besides, the energy characteristics of images at different noise levels can be used to estimate the noise energy distribution of the group matrix in the SVD domain [19]. As an enhancing step, iterative back projection is an effective way to suppress residual noise.…”

“…As an enhancing step, iterative back projection is an effective way to suppress residual noise. For improving the performance of the back projection process, a new noise standard deviation estimation method is given in [19], which results in a more effective denoising performance.…”