Hyperspectral Image Denoising via Subspace-Based Nonlocal Low-Rank and Sparse Factorization

Cao, Chunhong; Yu, Jie; Zhou, Chengyao; Hu, Kai; Xiao, Fen; Gao, Xieping

doi:10.1109/jstars.2019.2896031

Cited by 69 publications

(56 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Gaussian. When noise is a mixture of Gaussian noise, stripes, and impulse noise, the spectral subspace is usually estimated jointly with subspace coefficients of the HSI; for example, in double-factor-regularized low-rank tensor factorization (LRTF-DFR) [21] and subspace-based nonlocal low-rank and sparse factorization (SNLRSF) [22]. The joint estimation of the subspace and the corresponding coefficients of the HSI usually produce poor estimates of the subspace when HSI is affected by severe mixed noise.…”

Section: Related Workmentioning

confidence: 99%

“…The non-convex rank constraint is usually relaxed by minimizing the nuclear norm of HSIs, as done in the low-rank matrix recovery (LRMR) method [15] and in the weighted sum of weighted tensor nuclear norm minimization (WSWTNNM) method [16]. The low-rank structure of HSIs is also exploited by representing the spectral vectors of the clean image in an orthogonal subspace in [8,[17][18][19][20] for Gaussian noise removal and in [21,22] for mixed noise removal. Subspace representation is an explicit lowrank representation in the sense that the rank is constrained by the dimension of subspace.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Hyperspectral Image Mixed Noise Removal Using Subspace Representation and Deep CNN Image Prior

Zhuang

2021

Remote Sensing

View full text Add to dashboard Cite

The ever-increasing spectral resolution of hyperspectral images (HSIs) is often obtained at the cost of a decrease in the signal-to-noise ratio (SNR) of the measurements. The decreased SNR reduces the reliability of measured features or information extracted from HSIs, thus calling for effective denoising techniques. This work aims to estimate clean HSIs from observations corrupted by mixed noise (containing Gaussian noise, impulse noise, and dead-lines/stripes) by exploiting two main characteristics of hyperspectral data, namely low-rankness in the spectral domain and high correlation in the spatial domain. We take advantage of the spectral low-rankness of HSIs by representing spectral vectors in an orthogonal subspace, which is learned from observed images by a new method. Subspace representation coefficients of HSIs are learned by solving an optimization problem plugged with an image prior extracted from a neural denoising network. The proposed method is evaluated on simulated and real HSIs. An exhaustive array of experiments and comparisons with state-of-the-art denoisers were carried out.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Hyperspectral Image Mixed Noise Removal Using Subspace Representation and Deep CNN Image Prior

Zhuang

2021

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Recently, the image prior method based on nonlocal self-similarity [ 13 , 14 ] and low rank matrix approximating [ 15 , 16 , 17 , 18 ] can better preserve image edge details while denoising, which has achieved some success in image denoising [ 19 , 20 ]. Low rank matrix approximation aims to recover the underlying low rank matrix from degraded observations.…”

Section: Introductionmentioning

confidence: 99%

Weighted Schatten p-Norm Low Rank Error Constraint for Image Denoising

Cheng

2021

Entropy

View full text Add to dashboard Cite

Traditional image denoising algorithms obtain prior information from noisy images that are directly based on low rank matrix restoration, which pays little attention to the nonlocal self-similarity errors between clear images and noisy images. This paper proposes a new image denoising algorithm based on low rank matrix restoration in order to solve this problem. The proposed algorithm introduces the non-local self-similarity error between the clear image and noisy image into the weighted Schatten p-norm minimization model using the non-local self-similarity of the image. In addition, the low rank error is constrained by using Schatten p-norm to obtain a better low rank matrix in order to improve the performance of the image denoising algorithm. The results demonstrate that, on the classic data set, when comparing with block matching 3D filtering (BM3D), weighted nuclear norm minimization (WNNM), weighted Schatten p-norm minimization (WSNM), and FFDNet, the proposed algorithm achieves a higher peak signal-to-noise ratio, better denoising effect, and visual effects with improved robustness and generalization.

show abstract

“…However, in practice, a system of linear equations with matching unknowns are very rare and often ends up being sparse. The sparse solution of the equations also find numerous applications, especially in image processing [25], [9], [4], [17]. The field of sparse representation elicits its structure from the conventional transforms which ensures simplicity of processing, efficiency of representation, speed etc.…”

Section: Introductionmentioning

confidence: 99%

“…For a quantitative analysis of the obtained results, the algorithm uses PSNR, ISNR and SSIM and these values for cameraman, football and lena images each with dimensions 20 × 20, 40 × 40 and 60 × 60 are tabulated in the Tables(3)(4)(5). Due to space constraint, only tabulated results for Cameraman are presented.…”

mentioning

confidence: 99%

Analytics of Unconstrained Convex Problems in Sparse and Redundant Representations: A Case Study on Image Processing Applications

Patil¹,

Kulkarni²

2019

IJCA

View full text Add to dashboard Cite

The key objective of sparse and redundant representations is all about the introduction of a highly elegant data model with well-defined mathematical foundations. The structure of this model has been emerged from the conventional transforms represented in redundant unitary form. This draws out a new flavor of treatment to data. The appeal of this model is attributed to compact representation it facilitates. A lot of freedom persists in model adaptation to fit the data depending on the application.Many models are available aiming to serve the data of interest. Models can be given as mathematical descriptions or the conditions that the underlying signal of interest are believed to obey. Models that are available in image processing are Discrete Cosine Transform (DCT), Principal Component Analysis (PCA), wiener filtering, anisotropic diffusion, etc. Posing a new model requires a delicate attention to attain the simplest possible and reliable model while justifying the actual content of the data. To this Sparseland model emerges out as a new universal data model serving many applications in image processing.A class of applications in image processing demand the recovery of clean image from the naturally available perturbed images, which are treated as inverse problems. Some examples of such problems are image deblurring, image inpainting, etc. This paper discusses sparse approach to an inverse problem with image deblurring as a case study. The comparison of various shrinkage algorithms supporting sparse approach, used to serve this application is discussed along with the key constraints involved in these algorithms.

show abstract

Hyperspectral Image Denoising via Subspace-Based Nonlocal Low-Rank and Sparse Factorization

Cited by 69 publications

References 53 publications

Hyperspectral Image Mixed Noise Removal Using Subspace Representation and Deep CNN Image Prior

Hyperspectral Image Mixed Noise Removal Using Subspace Representation and Deep CNN Image Prior

Weighted Schatten p-Norm Low Rank Error Constraint for Image Denoising

Analytics of Unconstrained Convex Problems in Sparse and Redundant Representations: A Case Study on Image Processing Applications

Contact Info

Product

Resources

About