Self-Supervised Nonlinear Transform-Based Tensor Nuclear Norm for Multi-Dimensional Image Recovery

Luo, Yi-Si; Zhao, Xi-Le; Jiang, Tai-Xiang; Chang, Yi; Ng, Michael K.; Li, Chao

doi:10.1109/tip.2022.3176220

Cited by 46 publications

(7 citation statements)

References 70 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The sampling rates are set as {0.1, 0.2, 0.3}, and the standard deviation σ of Gaussian noise in the range of [0,1] is set as 0.1. The compared methods consist of seven model-based methods (i.e., TMac-TT [2], TNN [57], TRLRF [53], FTNN [15], TCTF [58], SN2TNN [23], and HLRTF [22]), two unsupervised deep learning-based methods (i.e., DIP2D [36] and DIP3D [36]), and two diffusion-based methods (i.e., DDRM [16] and DDNM [46]). HSI Denoising.…”

Section: Datasets and Compared Methodsmentioning

confidence: 99%

DDS2M: Self-Supervised Denoising Diffusion Spatio-Spectral Model for Hyperspectral Image Restoration

Miao¹,

Zhang²,

Zhang³

et al. 2023

Preprint

View full text Add to dashboard Cite

Diffusion models have recently received a surge of interest due to their impressive performance for image restoration, especially in terms of noise robustness. However, existing diffusion-based methods are trained on a large amount of training data and perform very well in-distribution, but can be quite susceptible to distribution shift. This is especially inappropriate for data-starved hyperspectral image (HSI) restoration. To tackle this problem, this work puts forth a self-supervised diffusion model for HSI restoration, namely Denoising Diffusion Spatio-Spectral Model (DDS2M), which works by inferring the parameters of the proposed Variational Spatio-Spectral Module (VS2M) during the reverse diffusion process, solely using the degraded HSI without any extra training data. In VS2M, a variational inference-based loss function is customized to enable the untrained spatial and spectral networks to learn the posterior distribution, which serves as the transitions of the sampling chain to help reverse the diffusion process. Benefiting from its self-supervised nature and the diffusion process, DDS2M enjoys stronger generalization ability to various HSIs compared to existing diffusion-based methods and superior robustness to noise compared to existing HSI restoration methods. Extensive experiments on HSI denoising, noisy HSI completion and super-resolution on a variety of HSIs demonstrate DDS2M's superiority over the existing task-specific state-of-the-arts.

show abstract

Section: Datasets and Compared Methodsmentioning

confidence: 99%

DDS2M: Self-Supervised Denoising Diffusion Spatio-Spectral Model for Hyperspectral Image Restoration

Miao¹,

Zhang²,

Zhang³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Besides that, the most similar works to TDNet are [47], [48], [49], but these works are mainly designed for hyperspectral image recovery tasks. In [47], the authors used deep generative networks for low-rank decomposition, and the method is robust to a wide range of distributions.…”

Section: Autoencoder Based Methodsmentioning

confidence: 99%

“…In [47], the authors used deep generative networks for low-rank decomposition, and the method is robust to a wide range of distributions. Luo et al [48] utilized a multilayer neural network to learn the nonlinear nature of multidimensional images, which is more effective than the linear transforms. Zhang et al [49] constructed the Kroneker bases by performing a maxpooling operation on the input features, and then the degraded hyperspectral image is recovered by a Conv layer with the spatial size of 3 × 3.…”

Section: Autoencoder Based Methodsmentioning

confidence: 99%

Tensor Decomposition-Inspired Convolutional Autoencoders for Hyperspectral Anomaly Detection

Sun

Zhao

Liu

et al. 2022

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Anomaly detection from hyperspectral images (HSI) is an important task in the remote sensing domain. Considering the three-order characteristics of HSI, many tensor decomposition based hyperspectral anomaly detection (HAD) models have been proposed and drawn much attention during the past decades. However, as most tensor decomposition based detectors are directly performed on the original HSI, the detection accuracy is usually limited due to the high-dimension and noise corruption of the HSI. Benefiting from the good capacity of autoencoders (AE) for feature extraction, in this paper, an enhanced tensor decomposition-inspired convolutional AE for HAD is proposed to address those problems, named TDNet. Within the proposed TD-Net, the traditional canonical-polyadic (CP) tensor decomposition model is innovatively alternated by a deep neural network (DNN), and the DNN tensor decomposition model performs more stably and robustly for noise. Specificly, a potential abnormal pixels remove strategy is firstly built to obtain the background training sets. Then, a DNN tensor decomposition-inspired convolutional AE is used to recover the original background information, which consists of an encoder, a low-rank tensor decomposition network, and a decoder. Finally, the residual errors between input HSI and recovered background are used for anomaly detection. Extensive experiments demonstrate the superiority of the TDNet in terms of both AUC values and ROC curves.

show abstract

“…For instance, Li et al [20] proposed a composite transform that combines a linear transform with a nonlinear counterpart, applying it to LRTC. Additionally, Luo et al [27] employed nonlinear transforms derived from nonlinear multilayer neural networks to represent the low-rank tensor. However, the approaches based on t-SVD mentioned above lack flexibility in capturing correlations between different modes of tensors, thus hindering the achievement of optimal tensor completion performance.…”

Section: Introductionmentioning

confidence: 99%

“…This approach effectively captures correlations between different modes, enhancing the low-rank features of the transformed tensor. Besides, 3DSTNN embeds a self-supervised non-linear transformation (S2NT) [27], into TNN. S2NT comprises multiple linear transformations and non-linear activation functions, resembling a nonlinear multilayer neural network.…”

Section: Introductionmentioning

confidence: 99%

Multi-Dimensional Image Recovery via Self-Supervised Nonlinear Transform Based a Three-Directional Tensor Nuclear Norm

Li,

Tu,

et al. 2024

NMTMA

View full text Add to dashboard Cite

Self-Supervised Nonlinear Transform-Based Tensor Nuclear Norm for Multi-Dimensional Image Recovery

Cited by 46 publications

References 70 publications

DDS2M: Self-Supervised Denoising Diffusion Spatio-Spectral Model for Hyperspectral Image Restoration

DDS2M: Self-Supervised Denoising Diffusion Spatio-Spectral Model for Hyperspectral Image Restoration

Tensor Decomposition-Inspired Convolutional Autoencoders for Hyperspectral Anomaly Detection

Multi-Dimensional Image Recovery via Self-Supervised Nonlinear Transform Based a Three-Directional Tensor Nuclear Norm

Contact Info

Product

Resources

About