Compressed Sensing Image Reconstruction Research Articles

Three-Dimensional Phase Resolved Functional Lung Magnetic Resonance Imaging.

Pulmonary magnetic resonance imaging (MRI) offers a variety of radiation-free techniques tailored to assess regional lung ventilation or its surrogates. These techniques encompass direct measurements, exemplified by hyperpolarized gas MRI and fluorinated gas MRI, as well as indirect measurements facilitated by oxygen-enhanced MRI and proton-based Fourier decomposition (FD) MRI. In recent times, there has been substantial progress in the field of FD MRI, which involved improving spatial/temporal resolution, refining sequence design and postprocessing, and developing a comprehensive whole-lung approach. The two-dimensional (2D) phase-resolved functional lung (PREFUL) MRI stands out as an FD-based approach developed for the comprehensive assessment of regional ventilation and perfusion dynamics, all within a single MR acquisition. Recently, a new advancement has been made with the development of 3D PREFUL to assess dynamic ventilation of the entire lung using 8 min exam with a self-gated sequence. The 3D PREFUL acquisition involves employing a stack-of-stars spoiled-gradient-echo sequence with a golden angle increment. Following the compressed sensing image reconstruction of approximately 40 breathing phases, all the reconstructed respiratory-resolved images undergo registration onto a fixed breathing phase. Subsequently, the ventilation parameters are extracted from the registered images. In a study cohort comprising healthy volunteers and patients with chronic obstructive pulmonary disease, the 3D PREFUL ventilation parameters demonstrated strong correlations with measurements obtained from pulmonary function tests. Additionally, the interscan repeatability of the 3D PREFUL technique was deemed to be acceptable, indicating its reliability for repeated assessments of the same individuals. In summary, 3D PREFUL ventilation MRI provides a whole lung coverage and captures ventilation dynamics with enhanced spatial resolution compared to 2D PREFUL. 3D PREFUL technique offers a cost-effective alternative to hyperpolarized 129Xe MRI, making it an attractive option for patient-friendly evaluation of pulmonary ventilation.

Multiple Complementary Priors for Multispectral Image Compressive Sensing Reconstruction.

Compressive sensing (CS) techniques using a few compressed measurements have drawn considerable interest in reconstructing multispectral imagery (MSI). Nonlocal-based tensor methods have been widely used for MSI-CS reconstruction, which employ the nonlocal self-similarity (NSS) property of MSI to obtain satisfactory results. However, such methods only consider the internal priors of MSI while ignoring important external image information, for example deep-driven priors learned from a corpus of natural image datasets. Meanwhile, they usually suffer from annoying ringing artifacts due to the aggregation of overlapping patches. In this article, we propose a novel approach for highly effective MSI-CS reconstruction using multiple complementary priors (MCPs). The proposed MCP jointly exploits nonlocal low-rank and deep image priors under a hybrid plug-and-play framework, which contains multiple pairs of complementary priors, namely, internal and external, shallow and deep, and NSS and local spatial priors. To make the optimization tractable, a well-known alternating direction method of multiplier (ADMM) algorithm based on the alternating minimization framework is developed to solve the proposed MCP-based MSI-CS reconstruction problem. Extensive experimental results demonstrate that the proposed MCP algorithm outperforms many state-of-the-art CS techniques in MSI reconstruction. The source code of the proposed MCP-based MSI-CS reconstruction algorithm is available at: https://github.com/zhazhiyuan/MCP_MSI_CS_Demo.git.

EIDNet: Extragradient-based iterative denoising network for image compressive sensing reconstruction

Compressive sensing (CS) is a kind of hardware-friendly technology in effective image reconstruction. Most of the existing CS reconstruction methods can be classified into model-driven iterative optimization methods and data-driven neural network methods. These two types of methods have their own problems with computational complexity and theoretical interpretability. Motivated by the deep unfolding theory, we develop the model-driven Extragradient-based Iterative Denoising Network, dubbed EIDNet, which is based on the denoising model of the iterative shrinkage threshold algorithm (ISTA). By unfolding the iterative denoising model into the deep learning network, EIDNet addresses the interpretability problem of neural network methods in CS reconstruction. Each denoising iteration of the optimization algorithm is mapped into one layer of the reconstruction network in EIDNet. The extragradient method is adopted to accelerate network convergence. In addition, we introduce a learning-based sampling and initialization network for adaptive sampling and initialization. Extensive experiments show that the proposed EIDNet can obtain better image CS reconstruction performance than other state-of-the-art CS methods while maintaining satisfactory reconstruction efficiency.

Multiply Complementary Priors for Image Compressive Sensing Reconstruction in Impulsive Noise

Impulsive noise is always present in real-world image Compressive Sensing (CS) acquisition systems, where existing CS reconstruction performance may seriously deteriorate. In this article, we propose a robust CS formulation for image reconstruction to suppress outliers in the presence of impulsive noise. To address this issue, we consider a novel truncated-Cauchy loss function as the metric of residual error to elevate the reconstruction robustness. Specifically, we design a complementary priors model to incorporate nonconvex nonlocal low-rank prior and deep denoiser prior for high-accuracy image reconstruction. By means of the half-quadratic optimization theory and generalized soft-thresholding technique, we also develop an alternative optimization algorithm for solving the induced nonconvex optimization problem. Numerical simulations demonstrate the robustness and accuracy of the proposed robust CS method compared to some recent CS methods for image reconstruction in impulsive noise.

Soft threshold iteration-based anti-noise compressed sensing image reconstruction network

Soft threshold iteration-based anti-noise compressed sensing image reconstruction network

Hyperspectral Compressive Image Reconstruction With Deep Tucker Decomposition and Spatial–Spectral Learning Network

Hyperspectral compressive imaging has taken advantage of compressive sensing theory to capture spectral information of the dynamic world in recent decades of years, where an optical encoder is employed to compress high dimensional signals into a single 2-D measurement. The core issue is how to reconstruct the underlying hyperspectral image (HSI), although deep neural network methods have achieved much success in compressed sensing image reconstruction in recent years, they still have some unsolved issues, such as tradeoffs between performance and efficiency, and accurate exploitation of cubic structure information. In this article, we propose a deep Tucker decomposition and spatial–spectral learning network (DS-net) to learn the tensor low-lank structure features and spatial–spectral correlation of HSI for reconstruction quality promotion. Inspired by tensor decomposition, we first construct a deep Tucker decomposition module to learn the principal components from different modes of the image features. Then, we cascade a series of decomposition modules to learn multihierarchical features. Furthermore, to jointly capture the spatial–spectral correlation of HSI, we propose a spatial–spectral correlation learning module in a U-net structure for more robust reconstruction performance. Finally, experimental results on both synthetic and real datasets demonstrate the superiority of the proposed method compared to several state-of-the-art methods in quantitative assessment and visual effects.

Image compressive sensing via hybrid regularization combining centralized group sparse representation and deep denoiser prior

To effectively solve the ill-posed image compressive sensing (CS) reconstruction problem, it is essential to properly exploit image prior knowledge. In this paper, we propose an efficient hybrid regularization approach for image CS reconstruction, which can simultaneously exploit both internal and external image priors in a unified framework. Specifically, a novel centralized group sparse representation (CGSR) model is designed to more effectively exploit internal image sparsity prior by suppressing the group sparse coding noise (GSCN), i.e., the difference between the group sparse coding coefficients of the observed image and those of the original image. Meanwhile, by taking advantage of the plug-and-play (PnP) image restoration framework, a state-of-the-art deep image denoiser is plugged into the optimization model of image CS reconstruction to implicitly exploit external deep denoiser prior. To make our hybrid internal and external image priors regularized image CS method (named as CGSR-D-CS) tractable and robust, an efficient algorithm based on the split Bregman iteration is developed to solve the optimization problem of CGSR-D-CS. Experimental results demonstrate that our CGSR-D-CS method outperforms some state-of-the-art image CS reconstruction methods (either model-based or deep learning-based methods) in terms of both objective quality and visual perception.

Utilizing the structure of a redundant dictionary comprised of wavelets and curvelets with compressed sensing

The discrete curvelet transform decomposes an image into a set of fundamental components that are distinguished by direction and size as well as a low-frequency representation. The curvelet representation is approximately sparse; thus, it is a useful sparsifying transformation to be used with compressed sensing. However, the low-frequency portion is seldom sparse. This manuscript presents a method to modify the redundant sparsifying transformation comprised of the wavelet and curvelet transforms to take advantage of this fact with compressed sensing image reconstruction. Instead of relying on sparsity for this low-frequency estimate, the Nyquist-Shannon theorem specifies a square region centered on the $0$ frequency to be collected, which is used to generate a blurry estimate. A Basis Pursuit Denoising problem is solved to determine the missing details after modifying the sparsifying transformation to take advantage of the known fully sampled region. Improvements in quality are shown on magnetic resonance and optical images.

Compressed Sensing Reconstruction Based on Residual Block and Dense Block

Existing compressed sensing image reconstruction algorithms have a large amount of computation and poor reconstruction quality. In order to solve the above problems, this paper proposes an image compressed sensing reconstruction network (RD-Net) based on residual blocks and dense blocks. The sampling network divides the input image into blocks with a block size of and uses convolution checks. These image blocks are compressed and sampled to obtain compressed sensing measurements. The initial reconstruction network uses Reshape+Concat to generate the initial reconstructed image. The deep reconstruction network is composed of residual blocks and dense blocks, which improves the quality of reconstructed images and also improves the running speed. Experimental results show that RD-Net improves the quality of reconstructed images.

Extra Proximal-Gradient Network with Learned Regularization for Image Compressive Sensing Reconstruction.

Learned optimization algorithms are promising approaches to inverse problems by leveraging advanced numerical optimization schemes and deep neural network techniques in machine learning. In this paper, we propose a novel deep neural network architecture imitating an extra proximal gradient algorithm to solve a general class of inverse problems with a focus on applications in image reconstruction. The proposed network features learned regularization that incorporates adaptive sparsification mappings, robust shrinkage selections, and nonlocal operators to improve solution quality. Numerical results demonstrate the improved efficiency and accuracy of the proposed network over several state-of-the-art methods on a variety of test problems.

A Review of Deep Learning Methods for Compressed Sensing Image Reconstruction and Its Medical Applications

Compressed sensing (CS) and its medical applications are active areas of research. In this paper, we review recent works using deep learning method to solve CS problem for images or medical imaging reconstruction including computed tomography (CT), magnetic resonance imaging (MRI) and positron-emission tomography (PET). We propose a novel framework to unify traditional iterative algorithms and deep learning approaches. In short, we define two projection operators toward image prior and data consistency, respectively, and any reconstruction algorithm can be decomposed to the two parts. Though deep learning methods can be divided into several categories, they all satisfies the framework. We built the relationship between different reconstruction methods of deep learning, and connect them to traditional methods through the proposed framework. It also indicates that the key to solve CS problem and its medical applications is how to depict the image prior. Based on the framework, we analyze the current deep learning methods and point out some important directions of research in the future.

Multiscale deep network for compressive sensing image reconstruction

Deep learning-based image compressive sensing methods have received extensive attention in recent years due to their superior learning ability and fast processing speed. The majority of existing image compressive sensing neural networks use single-scale sampling, whereas multiscale sampling has demonstrated excellent performance compared to single-scale. We propose a multiscale deep network for compressive sensing image reconstruction that consists of a multiscale sampling network and a reconstruction network. First, we use convolution to mimic the linear decomposition of images, and the convolution is learned during the training process. Then a sampling network captures compressive measurements across multiple decomposed scales. The reconstruction network, which includes both the initial and enhanced reconstruction networks, learns an end-to-end mapping between the compressed sensing (CS) measurements and the recovered images of the network. Experimental results indicate that the proposed network framework outperforms the existing CS methods in terms of objective metrics, peak signal to noise ratio (PSNR), structural similarity index, and subjective visual quality. Specifically, at a 0.1 sampling rate, using 10 images for testing, and the average PSNR maximum (minimum) gain is 5.95 dB (0.25 dB).

Group-Based Sparse Representation for Compressed Sensing Image Reconstruction with Joint Regularization

Achieving high-quality reconstructions of images is the focus of research in image compressed sensing. Group sparse representation improves the quality of reconstructed images by exploiting the non-local similarity of images; however, block-matching and dictionary learning in the image group construction process leads to a long reconstruction time and artifacts in the reconstructed images. To solve the above problems, a joint regularized image reconstruction model based on group sparse representation (GSR-JR) is proposed. A group sparse coefficients regularization term ensures the sparsity of the group coefficients and reduces the complexity of the model. The group sparse residual regularization term introduces the prior information of the image to improve the quality of the reconstructed image. The alternating direction multiplier method and iterative thresholding algorithm are applied to solve the optimization problem. Simulation experiments confirm that the optimized GSR-JR model is superior to other advanced image reconstruction models in reconstructed image quality and visual effects. When the sensing rate is 0.1, compared to the group sparse residual constraint with a nonlocal prior (GSRC-NLR) model, the gain of the peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) is up to 4.86 dB and 0.1189, respectively.

SWDGAN: GAN-based sampling and whole image denoising network for compressed sensing image reconstruction

We propose a framework named sampling and whole image denoising network based on generative adversarial network (GAN) for compressed sensing image reconstruction (SWDGAN) to reconstruct natural image from compressed sensing (CS) measurements. This work is devoted to balancing the performance of reconstruction quality, practicability, and running efficiency. Different from the recent deep learning reconstruction networks, we further enhance the feature representation and remove the blocking artifacts by introducing a whole image dense residual denoising module without affecting running efficiency. To improve the flexibility of sampling process and the practicability of our algorithm, a fully connected network without bias is applied in the sampling process, whose weight can be extracted separately and used as a measurement matrix. In this way, the measurements can be obtained by matrix multiplication in multiple running environments, not limited to deep learning framework. Besides, the sampling network can also improve the reconstruction quality even at low sampling ratios. Moreover, we remove batch normalization (BN) layer of reconstruction network to avoid the influence of BN artifact on reconstruction image. The experimental results illustrate that our method outperforms the most advanced traditional methods and deep learning-based methods in terms of reconstruction quality and running time.

Improving Compressed Sensing Image Reconstruction Based on Atmospheric Modulation Using the Distributed Cumulative Synthesis Method

The problem of long-distance imaging through time-varying scattering media, such as the atmosphere, is encountered in many science fields. Recent studies have demonstrated that random atmospheric variability can be considered a spatial light modulator in compressed sensing imaging. However, the quality of the reconstructed image needs to be further improved. In this paper, we propose a distributed cumulative synthesis method to improve the compressed sensing image reconstruction based on atmospheric modulation. For multiple original images of various types, the compressed sensing imaging simulation experiment with different sampling rates was conducted using the distributed cumulative synthesis method. The simulation results show that, compared with the imaging method using a single light source, the distributed cumulative synthesis method can effectively improve the quality of the reconstructed image, whether it is full sampling or undersampling. In addition, a sparsity impact factor is defined to quantify the reconstruction ability of the measurement matrix obtained by the distributed cumulative synthesis method. This value can be used as an evaluation index for the optimized design of the measurement matrix by the distributed cumulative synthesis method. Noise analysis shows that the proposed method has better anti-noise performance than the single light source imaging method.

Block compressed sensing image reconstruction via deep learning with smoothed projected Landweber

Compressed sensing (CS) is a signal processing framework for effectively reconstructing signal from a small number of measurements obtained by linear projections of the signal. It is an ongoing challenge for the real-time image reconstruction of the computational imaging, including single pixel imaging based on CS. We built a block-based CS (BCS) image reconstruction framework via a deep learning network with smoothed projected Landweber (SPL). A fully connected network performs both BCS linear sensing and non-linear reconstruction stages, and SPL removes the blocking artifacts due to incorporate Wiener filtering into projected Landweber (PL) method at each iteration. The sensing matrix and nonlinear prediction operator are jointly optimized, and the smoothing filtering is coalesced into the PL framework for eliminating high-frequency oscillatory blocking artifact. Experimental results reveal that the optimized scheme outperforms the approach only based on deep neural network. The reconstruction quality can be improved while being only slightly slower, especially the gain of structural similarity is significantly better than peak signal-to-noise ratio, and the reconstruction image texture details are vivid and natural. At 10% sensing rate, the structural similarity maximum (minimum) gain reaches 0.098 (0.021). The proposed approach is not only far superior to other state-of-the-art CS algorithms in terms of reconstruction time and quality but also comparable with up-to-date deep learning methods.

Nonlocal Low-Rank Regularization Combined with Bilateral Total Variation for Compressive Sensing Image Reconstruction

The use of non-local self-similarity prior between image blocks can improve image reconstruction performance significantly. We propose a compressive sensing image reconstruction algorithm that combines bilateral total variation and nonlocal low-rank regularization to overcome over-smoothing and degradation of edge information which result from the prior reconstructed image. The proposed algorithm makes use of the preservation of image edge information by bilateral total variation operator to enhance the edge details of the reconstructed image. In addition, we use weighted nuclear norm regularization as a low-rank constraint for similar blocks of the image. To solve this convex optimization problem, the Alternating Direction Method of Multipliers (ADMM) is employed to optimize and iterate the algorithm model effectively. Experimental results show that the proposed algorithm can obtain better image reconstruction quality than conventional algorithms with using total variation regularization or considering the nonlocal structure of the image only. At 10% sampling rate, the peak signal-to-noise ratio gain is up to 2.39 dB in noiseless measurements compared with Nonlocal Low-rank Regularization (NLR-CS). Reconstructed image comparison shows that the proposed algorithm retains more high frequency components. In noisy measurements, the proposed algorithm is robust to noise and the reconstructed image retains more detail information.

WITHDRAWN: Image Art Design Based on Image Compression Sensing System and Embedded Device

ABSTRACT The image Compressed Sensing (CS), two main challenges is to develop the design and reconstruction of the sample matrix. In one aspect, a random sample matrix is generally used and signal characteristics of the signal independent negligible. On the other hand, the most advanced art design image CS method made a very good performance. High computational complexity to meet two challenges recommend using convolution neuroimaging CS architecture. Network includes sampling network (called Convolutional Neural Network (CNN)) Joint Optimization of network reconstruction. Adaptation sampling learning image samples from a training matrix, which can accommodate a plurality of images of the configuration information, better reconstruction CS measurements. Reconstruction network, the network including linear and nonlinear initial reconstructed depth reconstruction networks, learning measuring terminal between CS and the reconstructed image to the end of the map. While achieving high speed of operation, the experimental results show that the most advanced reconstruction quality provided by the state. In addition, the results also show that the trained sample matrix can significantly improve the traditional CS image reconstruction method.

Hybrid High-Order and Fractional-Order Total Variation with Nonlocal Regularization for Compressive Sensing Image Reconstruction

Total variation often yields staircase artifacts in the smooth region of the image reconstruction. This paper proposes a hybrid high-order and fractional-order total variation with nonlocal regularization algorithm. The nonlocal means regularization is introduced to describe image structural prior information. By selecting appropriate weights in the fractional-order and high-order total variation coefficients, the proposed algorithm makes the fractional-order and the high-order total variation complement each other on image reconstruction. It can solve the problem of non-smooth in smooth areas when fractional-order total variation can enhance image edges and textures. In addition, it also addresses high-order total variation alleviates the staircase artifact produced by traditional total variation, still smooth the details of the image and the effect is not ideal. Meanwhile, the proposed algorithm suppresses painting-like effects caused by nonlocal means regularization. The Lagrange multiplier method and the alternating direction multipliers method are used to solve the regularization problem. By comparing with several state-of-the-art reconstruction algorithms, the proposed algorithm is more efficient. It does not only yield higher peak-signal-to-noise ratio (PSNR) and structural similarity (SSIM) but also retain abundant details and textures efficiently. When the measurement rate is 0.1, the gains of PSNR and SSIM are up to 1.896 dB and 0.048 dB respectively compared with total variation with nonlocal regularization (TV-NLR).

Image Reconstruction Based on Fused Features and Perceptual Loss Encoder-Decoder Residual Network for Space Optical Remote Sensing Images Compressive Sensing

Compressive sensing (CS) technology is introduced into space optical remote sensing image acquisition stage, which could make wireless image sensor network node quickly and accurately obtain images in the case of two constraints of limited battery power and expensive sensor costs. On this basis, in order to further improve the quality of CS image reconstruction, we propose fused features and perceptual loss encoder-decoder residual network (FFPL-EDRNet) for image reconstruction. FFPL-EDRNet consists of a convolution layer and a reconstruction network. We train FFPL-EDRNet end-to-end, thus greatly simplifying the pre-processing and post-processing process and eliminating the block effect of reconstructed images. The reconstruction network is based on residual network, which introduces multi-scale feature extraction, multi-scale feature combination and multi-level feature combination. Feature fusion integrates low-level information with high-level information to reduce reconstruction error. The perceptual loss function based on pretrained InceptionV3 uses the weighted mean square error to define the loss value between the reconstructed image feature and the label image feature, which makes the reconstructed image more semantically similar to label image. In the measurement procedure, we use convolution to achieve block compression measurement, so as to obtain full image measurements. For image reconstruction, we firstly use a deconvolution layer to initially reconstruct the image and then use the residual network to refine the initial reconstructed image. The experimental results show that: in the case of measurement rates (MRs) of 0.25, 0.10, 0.04 and 0.01, the peak signal-to-noise ratio (PSNR) = 27.502, 26.804, 24.593, 21.359 and structural similarity (SSIM) = 0.842, 0.816, 0.720, 0.568 of the reconstructed images obtained by FFPL-EDRNet. Therefore, Our FFPL-EDRNet could enhance the quality of image reconstruction.