- 标题
- 摘要
- 关键词
- 实验方案
- 产品
-
Low-rank Bayesian tensor factorization for hyperspectral image denoising
摘要： In this paper, we present a low-rank Bayesian tensor factorization approach for hyperspectral image (HSI) denoising problem, where zero-mean white and homogeneous Gaussian additive noise is removed from a given HSI. The approach is based on two intrinsic properties underlying a HSI, i.e., the global correlation along spectrum (GCS) and nonlocal self-similarity across space (NSS). We first adaptively construct the patch-based tensor representation for the HSI to extract the NSS knowledge while preserving the property of GCS. Then, we employ the low rank property in this representation to design a hierarchical probabilistic model based on Bayesian tensor factorization to capture the inherent spatial-spectral correlation of HSI, which can be effectively solved under the variational Bayesian framework. Furthermore, through incorporating these two procedures in an iterative manner, we build an effective HSI denoising model to recover HSI from its corruption. This leads to a state-of-the-art denoising performance, consistently surpassing recently published leading HSI denoising methods in terms of both comprehensive quantitative assessments and subjective visual quality.
关键词： Hyperspectral image denoising,Global correlation along spectrum,Full Bayesian CP factorization,Nonlocal self-similarity,Variational Bayesian inference,Tensor rank auto determination
更新于2025-09-23 15:23:52
-
[IEEE ICASSP 2018 - 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) - Calgary, AB (2018.4.15-2018.4.20)] 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) - Group Sparsity Residual with Non-Local Samples for Image Denoising
摘要： Inspired by group-based sparse coding, recently proposed group sparsity residual (GSR) scheme has demonstrated superior performance in image processing. However, one challenge in GSR is to estimate the residual by using a proper reference of the group-based sparse coding (GSC), which is desired to be as close to the truth as possible. Previous researches utilized the estimations from other algorithms (i.e., GMM or BM3D), which are either not accurate or too slow. In this paper, we propose to use the Non-Local Samples (NLS) as reference in the GSR regime for image denoising, thus termed GSR-NLS. More specifically, we first obtain a good estimation of the group sparse coefficients by the image non-local self-similarity, and then solve the GSR model by an effective iterative shrinkage algorithm. Experimental results demonstrate that the proposed GSR-NLS not only outperforms many state-of-the-art methods, but also delivers the competitive advantage of speed.
关键词： Image denoising,group sparsity residual,iterative shrinkage algorithm,group-based sparse coding,non-local self-similarity
更新于2025-09-23 15:23:52
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Adaptive Hyperspectral Mixed Noise Removal
摘要： This paper proposes a new denoising method for hyperspectral images (HSIs) corrupted by mixtures (in a statistical sense) of stripe noise, Gaussian noise, and impulsive noise. The proposed method has three distinctive features: 1) it exploits the intrinsic characteristics of HSIs, namely, low-rank and self-similarity; 2) the observation noise is assumed to be additive and modeled by a mixture of Gaussian (MoG) densities; 3) the inference is performed with an expectation maximization (EM) algorithm, which, in addition to the clean HSI, also estimates the mixture parameters (posterior probability of each mode and variances). Comparisons of the proposed method with state-of-the-art algorithms provide experimental evidence of the effectiveness of the proposed denoising algorithm.
关键词： expectation maximization,low-rank,mixed noise,self-similarity,Denoising,mixture of Gaussians,hyperspectral images
更新于2025-09-23 15:22:29
-
[IEEE 2018 5th International Conference on Systems and Informatics (ICSAI) - Nanjing, China (2018.11.10-2018.11.12)] 2018 5th International Conference on Systems and Informatics (ICSAI) - An Enhanced Lowrank Algorithm for Image Denoising
摘要： There are great breakthroughs in image denoising based on the introduction of sparse representation and low rank theory. Some techniques, including BM3D and SAIST are brought forward and applied to various vision tasks. In this paper, we propose an enhanced SAIST algorithm for image denoising. These improvements are mainly implemented in the following aspects. First, when matching similar blocks, matching results are depended on block distances which affected by noise interference. Thus DCT pre-filtering is introduced before aggregation because it can effectively suppress measurement errors of block distances. Second, the relevance of image patches which affects the singular value thresholding is not considered in sample mean. So a weighted sample mean calculation method is proposed to make the singular value thresholding more adaptive. The experimental results show that this improved algorithm achieves a better performance than the original algorithm in terms of both objective criterion and subjective visual quality.
关键词： self-similarity,pre-filtering,singular value thresholding,low-rank method
更新于2025-09-23 15:22:29
-
[Lecture Notes in Electrical Engineering] Proceedings of 2018 Chinese Intelligent Systems Conference Volume 528 (Volume I) || Weighted Tensor Schatten p-norm Minimization for Image Denoising
摘要： In the traditional non-local similar patches based denoising algorithms, the image patches are ?rstly ?atted into a vector, which ignores the spatial layout information within the image patches that can be used for improving the denoising performance. To deal with this issue, we propose a weighted tensor Schatten p-norm minimization (WTSN) algorithm for image denoising and use alternating direction method (ADM) to solve it. In WTSN, the image patches are treated as matrix instead of vectorizing them, and thus make full use of information within the structure of the image patches. Furthermore, the employed Schatten p-norm requires much weaker incoherence conditions and can ?nd sparser solutions than the nuclear norm, and thus is more robust against noise and outliers. Experimental results show that the proposed WTSN algorithm outperforms many state-of-the-art denoising algorithms in terms of both quantitative measure and visual perception quality.
关键词： WTSN,Alternating direction method,Nonlocal self-similarity,Image denoising
更新于2025-09-23 15:21:21
-
[IEEE 2019 IEEE International Conference on Manipulation, Manufacturing and Measurement on the Nanoscale (3M-NANO) - Zhenjiang, China (2019.8.4-2019.8.8)] 2019 IEEE International Conference on Manipulation, Manufacturing and Measurement on the Nanoscale (3M-NANO) - Modification of Wettability Property of NITI Alloy by Laser Texturing and Carbon Ion Implantation
摘要： Nonlocal self-similarity of images has attracted considerable interest in the field of image processing and has led to several state-of-the-art image denoising algorithms, such as block matching and 3-D, principal component analysis with local pixel grouping, patch-based locally optimal wiener, and spatially adaptive iterative singular-value thresholding. In this paper, we propose a computationally simple denoising algorithm using the nonlocal self-similarity and the low-rank approximation (LRA). The proposed method consists of three basic steps. First, our method classifies similar image patches by the block-matching technique to form the similar patch groups, which results in the similar patch groups to be low rank. Next, each group of similar patches is factorized by singular value decomposition (SVD) and estimated by taking only a few largest singular values and corresponding singular vectors. Finally, an initial denoised image is generated by aggregating all processed patches. For low-rank matrices, SVD can provide the optimal energy compaction in the least square sense. The proposed method exploits the optimal energy compaction property of SVD to lead an LRA of similar patch groups. Unlike other SVD-based methods, the LRA in SVD domain avoids learning the local basis for representing image patches, which usually is computationally expensive. The experimental results demonstrate that the proposed method can effectively reduce noise and be competitive with the current state-of-the-art denoising algorithms in terms of both quantitative metrics and subjective visual quality.
关键词： self-similarity,Back projection,patch grouping,image denoising,low-rank approximation (LRA),singular value decomposition (SVD)
更新于2025-09-23 15:21:01
-
[IEEE 2019 Days on Diffraction (DD) - St. Petersburg, Russia (2019.6.3-2019.6.7)] 2019 Days on Diffraction (DD) - Novel types of mode dispersion of optical vortices in twisted optical fibers
摘要： Nonlocal self-similarity of images has attracted considerable interest in the field of image processing and has led to several state-of-the-art image denoising algorithms, such as block matching and 3-D, principal component analysis with local pixel grouping, patch-based locally optimal wiener, and spatially adaptive iterative singular-value thresholding. In this paper, we propose a computationally simple denoising algorithm using the nonlocal self-similarity and the low-rank approximation (LRA). The proposed method consists of three basic steps. First, our method classifies similar image patches by the block-matching technique to form the similar patch groups, which results in the similar patch groups to be low rank. Next, each group of similar patches is factorized by singular value decomposition (SVD) and estimated by taking only a few largest singular values and corresponding singular vectors. Finally, an initial denoised image is generated by aggregating all processed patches. For low-rank matrices, SVD can provide the optimal energy compaction in the least square sense. The proposed method exploits the optimal energy compaction property of SVD to lead an LRA of similar patch groups. Unlike other SVD-based methods, the LRA in SVD domain avoids learning the local basis for representing image patches, which usually is computationally expensive. The experimental results demonstrate that the proposed method can effectively reduce noise and be competitive with the current state-of-the-art denoising algorithms in terms of both quantitative metrics and subjective visual quality.
关键词： patch grouping,Back projection,low-rank approximation (LRA),singular value decomposition (SVD),image denoising,self-similarity
更新于2025-09-19 17:13:59
-
[IEEE 2019 Conference on Lasers and Electro-Optics Europe & European Quantum Electronics Conference (CLEO/Europe-EQEC) - Munich, Germany (2019.6.23-2019.6.27)] 2019 Conference on Lasers and Electro-Optics Europe & European Quantum Electronics Conference (CLEO/Europe-EQEC) - Electromagnetic Diffraction by Fractal Dusts, Triangles and Carpets: A Kirchhoff Approach to Circulation
摘要： The diffraction of plane waves by perfectly-conducting thin screens is of fundamental physical and mathematical interest in electromagnetics [1]. Classic laser-optics experiments involve both open (single- and double-slit arrangements) and closed (circular and regular-polygon) apertures, with analyses often being confined to the Fresnel [2] and Fraunhofer [3] regimes. Here, we consider a class of scattering problem involving fully-2D fractal screens, where the scatterer possesses the property of self-similarity. A more general formulation of the diffracted wave, based on Kirchhoff’s theory and 3D Green’s functions [4], is also deployed.
关键词： Kirchhoff’s theory,self-similarity,diffraction,Green’s functions,fractal screens
更新于2025-09-16 10:30:52
-
[IEEE IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Valencia, Spain (2018.7.22-2018.7.27)] IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium - Hyperspectral Imagery Denoising Using Multi-Linear Weighted Nuclear Norm Minimization
摘要： Classical matrix-based denoising methods for hyperspectral imagery (HSI) may cause spatial and spectral distortion. To improve denoising performance, a multi-linear weighted nuclear norm minimization was proposed for HSI denoising. By considering spectral continuity and inter-dependency of three unfolding modes, a multi-linear rank was proposed to model the spatial and spectral nonlocal similarity. To make the proposed method more tractable, a variable splitting based technique was used to solve the optimization problem. Experiment results reveal that the proposed method outperforms state-of-the-art methods both visually and quantitatively.
关键词： multi-linear rank,multi-linear weighted nuclear norm,nonlocal self-similarity,hyperspectral imagery denoising
更新于2025-09-10 09:29:36
-
[IEEE 2018 International Symposium in Sensing and Instrumentation in IoT Era (ISSI) - Shanghai, China (2018.9.6-2018.9.7)] 2018 International Symposium in Sensing and Instrumentation in IoT Era (ISSI) - An Expected Patch Log Likelihood Denoising Method Based on Internal and External Image Similarity
摘要： Natural images always exhibit a certain nonlocal self-similarity property, which implies that the patch matrix formed by similar image patches is low-rank. In this paper, the self-similarity of images is combined with the EPLL (Expected patch log likelihood) method based on external similarity, and an EPLL denoising model based on internal and external image similarity is proposed. The experimental results show that compared with the original EPLL method, the proposed method not only has higher quantization index, but also has a good visual effect.
关键词： image denoising,low-rank,self-similarity,expected patch log likelihood
更新于2025-09-10 09:29:36